OpenRig

CLI reference

Generated from OpenRig 0.5.14 (cc75efdd), commit cc75efdd17fb967bde7cff6c5805791986af78d8.

Commands and flags below come from installed help. A flag belongs to the command where it is listed; parent/global options are listed separately. Hidden commands and synthetic help subcommands are outside this index.

Command Purpose
rig add Add a member to an existing pod in a running rig
rig adopt Materialize topology and bind discovered live sessions
rig agent Manage agent specs
rig agent-image Browse, snapshot, and manage agent images (PL-016)
rig agent-image create Capture a productive seat's resumable state into a new agent image
rig agent-image delete Delete an agent image (subject to evidence-preservation guard)
rig agent-image list List all agent images in the library
rig agent-image pin Pin an image so prune cannot delete it
rig agent-image preview Show manifest + sized supplementary file metadata + starter snippet
rig agent-image prune Delete evictable images (protected by evidence-preservation guard)
rig agent-image show Show image manifest + statistics
rig agent-image sync Re-walk discovery roots and refresh the library index
rig agent-image unpin Unpin an image
rig agent validate Validate an agent spec (agent.yaml)
rig archive Archive a rig (soft + reversible: hides it from the default view, retains all data). NOT a delete; reverse with 'rig unarchive'.
rig ask Search rig transcript history with a natural language question
rig attach Attach the current shell or agent into a rig node
rig auth Manage agent auth profiles (CLI-local; runtime via --runtime). Tokens are never printed, logged, or stored.
rig auth list List saved profiles by name.
rig auth save Snapshot the active auth state into a named profile (file copy; contents never echoed).
rig auth seats Seat -> profile registry (metadata only; not proof of a live account).
rig auth seats list List seat -> profile mappings.
rig auth seats report Counts: total / known / unknown / malformed.
rig auth seats set Upsert a seat -> profile metadata row.
rig auth seats show Show the registry row for one seat.
rig auth status Auth-file presence + login state (no secrets).
rig auth switch Activate a saved profile.
rig auth validate Check a profile's file mode + JSON parseability (NOT live-auth).
rig bind Bind a discovered session to a rig node (existing or new)
rig bootstrap Bootstrap a rig from a spec file
rig broadcast Send a message to multiple agent sessions
rig bundle Manage rig bundles
rig bundle create Create a .rigbundle from a rig spec
rig bundle history List bundle install audit records from ~/.openrig/bundle-audit.jsonl
rig bundle inspect Inspect a .rigbundle
rig bundle install Install a .rigbundle (bootstrap from bundle)
rig capture Capture terminal output from agent sessions
rig chatroom Chat room for rig communication
rig chatroom clear No description supplied by help.
rig chatroom history No description supplied by help.
rig chatroom send No description supplied by help.
rig chatroom topic No description supplied by help.
rig chatroom wait No description supplied by help.
rig chatroom watch No description supplied by help.
rig compact Manually run the guided compaction sequence (prep → /compact → restore → audit) for one Claude seat
rig compact-plan Plan Claude compact-in-place candidates without compacting anything
rig config Inspect and change OpenRig configuration
rig config get Read a single config value
rig config init-workspace Scaffold a repo-ready project workspace
rig config reset Clear a config override (or delete the entire file when no key given)
rig config set Set a config value
rig context Browse, preview, compose, and manage operator-authored context packs
rig context add Install a pack; --git discovers a pack in a Git repository and retains its update relationship
rig context compose Compose ordered files into a durable context-pack ref (never delivers)
rig context get Serve the assembled bundle for an agent to load on demand (the pull verb)
rig context list List all context packs in the library
rig context preview Show the assembled bundle without delivering it
rig context profile No description supplied by help.
rig context recap-write Write this seat's authored RECAP (decisions-with-rationale) at the handover boundary; supersedes the previous recap into the seat's chain
rig context rm Remove a context pack from the library by its path-like ref
rig context show Show pack manifest + per-file metadata
rig context source Inspect Git checkout vs served context, or explicitly fetch/merge and select an update
rig context source inspect Read local revision, edits, conflicts and selection; no fetch or consumption claim
rig context source update Explicitly fetch/merge the upstream, then select the clean pack; refuse local selection edits
rig context sync Re-walk discovery roots and refresh the library index
rig context trace Walk the topology tree for one chain filename (instance -> rig -> optional pod -> optional seat), keyed off topology.root
rig context work-install Resolve System World plus project work context and the managed skill loadout
rig crash-cart Emit the daemon-down recovery verdict + discovery as JSON (read-only).
rig create Create a one-seat rig without writing a spec
rig daemon Manage the OpenRig daemon
rig daemon logs Show daemon logs
rig daemon start Start the daemon
rig daemon status Show daemon status
rig daemon stop Stop the daemon (10s shutdown budget; 12s process wait; incomplete drain exits nonzero)
rig destroy Destroy OpenRig local state for recovery
rig discover Scan for unmanaged tmux sessions
rig doctor Verify OpenRig install health
rig down Tear down a rig
rig env Inspect and control rig environment services for service-backed rigs and managed apps
rig env down No description supplied by help.
rig env logs No description supplied by help.
rig env status No description supplied by help.
rig expand Add a pod to a running rig
rig export Export a rig spec as YAML
rig file Cross-host file movement over ssh/rsync (v0: one explicit verb — copy)
rig file copy Copy one file. <hostId>:<absolute-path> = remote (ssh registry entry); bare path = local; local files with a colon need the ./ prefix. Valid shapes: local→remote, remote→local, local→local. An existing destination is OVERWRITTEN — preview with --dry-run.
rig fork Fork a live seat's context into a new seat (composes the shipped agent-image fork path)
rig gateway Gateway: the human registry + connector surfaces
rig gateway human Manage human specs (file-per-human fragments under gateway/humans/)
rig gateway human add Add a human fragment (verb-add-only; the fragment is truth, the registry is a GENERATED projection)
rig gateway human list Show the configured human (single-human surface per A1/R5; several fragments render honestly with a 0.5.7 advisory)
rig gateway human remove Remove a human: refuses while open conversations or non-terminal rows exist (--force archives and records what was orphaned; the fragment is archived, never deleted)
rig gateway human set Edit one field through the verb (same validation as add; re-projection immediate). Fields: display-name, delivery-class, away, binding.<n>
rig gateway human show Show the effective record: fragment values + which defaults filled the rest, with provenance
rig grow Add one or more seats to a running rig
rig handover Hand a seat to a successor: create -> deliver context -> verify continuity -> rebind
rig health Inspect read-only, explainable system health records
rig health checkpoint Inspect or submit an outcome-boundary lineage census (not a per-edit ritual)
rig health diagnose Preview policy admission; --apply creates or re-presents bounded diagnostic context
rig health diagnosis Read occurrences and record agent-owned dispositions
rig health diagnosis list List occurrence summaries; evidence payloads require --full
rig health diagnosis notify Explicitly request human delivery under policy and verified connector readiness
rig health diagnosis record No description supplied by help.
rig health diagnosis show Inspect current state and decisions; expand retained evidence deliberately
rig health explain Explain one health finding from its canonical bounded record
rig health policy Inspect effective policy and engine state; apply edited JSON with --file
rig heartbeat Show workflow execution proof state from queue files
rig host Manage the multi-host registry (~/.openrig/hosts.yaml)
rig host add Add a host entry (validated with the registry loader's own rules)
rig host doctor Stepwise host verification (+ --posture for the product-factory-vps baseline)
rig host list List registered hosts w/ status + selected marker (config pointers only — never secret values)
rig host pair Pair with a remote host from one pasted address — one approval on the target, done (FR-6)
rig host rename Rename THIS host's display name (renders in dashboard, explorer, ls, whoami)
rig host select Select which host you are viewing/acting on (persisted; 'local' returns to this host)
rig import Import a rig spec from YAML
rig launch Launch or relaunch a node in a running rig
rig mcp MCP server for agent integration
rig mcp serve Start MCP server (stdio transport)
rig mode Inspect and explicitly set scoped operating posture (human-led/delegated) and operator-context modes.
rig mode cite Emit a citation line for the effective mode at a read context. Per convention §Citation Rules.
rig mode defaults Print the recommended per-mode 8×7 + default-scope + stale rule.
rig mode effective Inspect effective operating posture, scope, phase and source. Resolved unset scopes default to human-led; unreadable or ambiguous identity stays unknown.
rig mode set Propose a mode binding. Without --confirm, restates the proposed binding and exits 2 (no daemon write). With --confirm, sets it.
rig mode show List all operator-context-mode bindings.
rig mode unset Delete one binding (operator-only). Scope: global_host | rig | project | mission | workstream | qitem.
rig package Manage agent packages (legacy)
rig package install Install a package
rig package list List installed packages
rig package plan Preview install plan (dry run)
rig package rollback Rollback an install
rig package validate Validate a package manifest
rig parked Are we parked? Derived diagnosis: stopped seats owing work; HELD is healthy only while its recorded wake is live
rig plugin Inspect plugins (read-only)
rig plugin list List discoverable plugins (aggregated across vendored + runtime caches)
rig plugin show Show plugin manifest + skills + hooks + mcp servers
rig plugin used-by List agents referencing this plugin in their profile.uses.plugins[]
rig plugin validate Validate plugin manifest + skill frontmatter against agentskills.io spec
rig policy Teach and record the rig-level permission policy. OpenRig bakes NO allow/ask/deny permission policy — the harness-native permissions are the control surface. OpenRig records posture into RigSpec; harness-native permissions enforce — never runtime enforcement. (The context-mode verb formerly at this name is now: rig mode.)
rig policy apply Record a policy choice into an EXISTING rig spec (the same recording flow as rig setup --policy, which stays as the setup-step composition). OpenRig bakes NO allow/ask/deny permission policy — the harness-native permissions are the control surface. OpenRig records posture into RigSpec; harness-native permissions enforce — never runtime enforcement.
rig policy current Show the EFFECTIVE permission policy for a rig spec and what would apply, per site (rig + member overrides), through the authoritative validator/resolver.
rig policy list List the built-in permission-policy templates, plus the custom policies visible in a spec context. OpenRig bakes NO allow/ask/deny permission policy — the harness-native permissions are the control surface. OpenRig records posture into RigSpec; harness-native permissions enforce — never runtime enforcement.
rig policy show Show one built-in policy choice, or validate + open a CUSTOM policy spec by ref (relative to --spec's directory, else cwd).
rig preflight Check system readiness for OpenRig
rig project Coordination L2 — agent-backed classifier with daemon-enforced lease + idempotency + reclaim
rig project classify Project a stream item with classification fields (idempotent on stream_item_id; requires active lease)
rig project lease-acquire Acquire the active classifier lease for the caller
rig project lease-heartbeat Send a heartbeat for an active classifier lease (extends TTL)
rig project lease-show Show the currently-active classifier lease
rig project list List project classifications with filters
rig project reclaim-classifier Operator-verb: reclaim the active classifier lease. Use --if-dead to refuse if holder is still alive.
rig project show Show one project classification
rig proof Capture evidence (add), record an attributed item judgment (judge), and read derived readiness (show). Capture, policy acceptance, higher outcome judgment and publication are separate.
rig proof add Drop a proof artifact: authors the C1 frontmatter from flags, writes <slice>/proof/<name>, echoes the parsed header. Contract/self-check/C8 outputs are advisories (exit 0) — never gates.
rig proof judge Record one attributed item judgment and derive readiness. Use mission/slices/slice#item (index, text or ID). Policy inherits proofPolicy.judges from owning slice, mission or project.
rig proof show Read current attributed proof readiness for a slice, mission or active project; no status files are changed.
rig provider Provider accounts, usage signals, and interruption-safe account switching
rig provider accounts The accounts block of the provider read model
rig provider bindings The bindings block of the provider read model
rig provider precheck Whether switching a seat to an account is safe (never offers an unsafe switch)
rig provider signals The signals block of the provider read model
rig provider status The whole four-block provider read model (accounts, bindings, signals)
rig provider switch Switch a seat to an account (precheck-gated; the daemon orchestrates the switch)
rig ps List rigs and their status
rig queue Coordination L3 — owned-work queue + inbox/outbox
rig queue block Park a qitem as HELD with a continuation and wake. Choose a watchdog id, timer, or live blocker.
rig queue claim Claim a qitem (pending → in-progress); computes closure_required_at from tier
rig queue create Create a new qitem
rig queue fallback Reroute a qitem to a fallback destination (e.g. unreachable seat)
rig queue handoff Transactional handoff: closes source as handed-off + creates new qitem owned by --to
rig queue handoff-and-complete Atomic close (state=done, closure_reason=handed_off_to) + create new qitem owned by --to. Variant of handoff that fully terminates the source qitem.
rig queue inbox-absorb Absorb a pending inbox entry into the receiver's main queue
rig queue inbox-deny Deny a pending inbox entry with a recorded reason
rig queue inbox-drop Drop a mailbox-style entry into a destination's inbox
rig queue inbox-pending List pending inbox entries for a destination seat
rig queue list List qitems (default: active + compact + current-rig; like 'docker ps')
rig queue outbox-list List outbox entries for a sender seat
rig queue outbox-record Record an outbound dispatch in the sender's outbox
rig queue overdue List in-progress qitems past their closure_required_at deadline (current rig, bounded, body-free by default)
rig queue resolve Resolve a leg-1 parked qitem (state=blocked on a human seat): records the decision text durably in queue_transitions, unparks blocked -> in-progress, and nudges the owner. Non-closure.
rig queue show Show one qitem and its derived waiting state (bounded preview; --full for complete body)
rig queue transitions Show the append-only transition log for a qitem
rig queue unclaim Release a claimed qitem (in-progress → pending)
rig queue undelivered List PENDING qitems whose create-path nudge FAILED (delivery never reached the destination; current rig, bounded, body-free by default)
rig queue update Append a note and/or mutate qitem state. A note without --state never changes row state. state=done REQUIRES --closure-reason (one of: handed_off_to, blocked_on, denied, canceled, no-follow-on, escalation). Closure ≠ acceptance: handed_off_to records delivery to the next stage; acceptance is the next stage's verdict on its own qitem, not this closure.
rig queue whoami Show the caller's queue position from the daemon's perspective
rig reconcile-session Adopt a live, hand-resumed session back into its persisted node (no launch, no input)
rig release Release claimed sessions from a rig without killing tmux sessions
rig remove Remove a node from a running rig
rig requirements Check requirements for a rig spec
rig restore Restore a rig from a snapshot
rig restore-check Check restore readiness across running rigs
rig restore-packet Generate, read, and validate cross-runtime restore packets per the v0 standard
rig restore-packet read Render a restore packet contents (human or --json)
rig restore-packet validate Validate a restore packet against the v0 schema
rig restore-packet write Generate a restore packet from a source session or JSONL file
rig restore status Show the derived current receipt for one restore attempt
rig scope Scope tree primitive: missions, slices, sub-slices (per conventions/scope-and-versioning)
rig scope audit Read-only scope audit: flag scope findings and show the advisory dependency graph
rig scope mission Mission-tier commands
rig scope mission approve Approve a mission: same staged/delegated semantics as slice approve, at mission tier.
rig scope mission create Create a new mission with SPEC.md and mission.yaml (mints a stable dot-ID into frontmatter)
rig scope mission graph Show advisory sibling build-order edges and the current ready set
rig scope mission ls List missions (top-level folders with SPEC.md or a legacy README.md)
rig scope mission progress Update a mission's progress rail deterministically (append a row, or set a row's status)
rig scope mission repair Backfill missing PROGRESS.md + conform mandatory frontmatter (id/stage/verified) for a mission and its slices; idempotent
rig scope mission show Inspect a single mission
rig scope mission stage Set a mission's epistemic stage (wip | provisional | established | canonical | superseded | retired); superseded needs --successor
rig scope mission verified Stamp a mission's verified line: <today> against <source> (provenance mandatory; overwrites the prior line)
rig scope resolve-notes Resolve the readable mission notes file for an absolute work-node directory
rig scope slice Slice-tier commands
rig scope slice approve Approve a slice: writes the frontmatter stamp + an append-only audit row (daemon-side, one operation). --scope spec = the PLAN-LOCK (PRD-matches-intent; this artifact set gets built); delivery (default) = the PROOF-LOCK (terminal sign-off). Approval is freeze/sign-off, never proven-green. Conventions SSOT: docs/reference/sdlc-conventions.md (installed: $OPENRIG_HOME/reference/sdlc-conventions.md).
rig scope slice close Close a slice (move to <mission>/closed/, update status)
rig scope slice create Create a new slice with SPEC.md, slice.yaml, PROGRESS.md, PROOF.md, and proof/. Conventions SSOT: docs/reference/sdlc-conventions.md (installed: $OPENRIG_HOME/reference/sdlc-conventions.md).
rig scope slice ls List slices in a mission (or across all missions)
rig scope slice move Move a slice between missions (re-numbers in destination)
rig scope slice progress Update a slice's progress rail deterministically (append a row, or set a row's status)
rig scope slice repair Backfill a slice's missing PROGRESS.md + conform mandatory frontmatter (id/stage/verified); idempotent
rig scope slice ship Ship a slice to a release mission (preserves git history)
rig scope slice show Inspect a single slice (frontmatter + README + children)
rig scope slice stage Set a slice's epistemic stage (wip | provisional | established | canonical | superseded | retired); superseded needs --successor
rig scope slice verified Stamp a slice's verified line: <today> against <source> (provenance mandatory; overwrites the prior line)
rig seat Inspect OpenRig seat observability state
rig seat clean Return a DEAD seat to launchable (clears stale binding + session records; audited)
rig seat clear-attention Clear stuck attention_required startup status with evidence or operator attestation
rig seat handover Plan a safe two-phase seat handover
rig seat launch Launch a deliberate fresh occupant for exactly one existing seat
rig seat set-model Persist a seat's model id (audited); subsequent managed resumes use the new model
rig seat set-resume-token Set a managed seat's durable resume token (token read from stdin; attested + audited)
rig seat status Show read-only seat handover observability status
rig seat stop Stop exactly one LIVE managed seat (kills only that seat's tmux session; audited)
rig seat switch-client Retarget an attached tmux client's view to the seat's canonical session (view-only)
rig send Send a message to an agent's terminal
rig setup Prepare the machine for OpenRig
rig shrink Remove an entire pod from a running rig
rig skill Skill management and audit
rig skill audit Read-only skill provenance and freshness audit
rig skill loadout Inspect or reconcile the composed managed skill loadout for one Claude/Codex working directory
rig slack Slack connector: configuration + in-daemon subsystem admin (S10)
rig slack disable Disable the connector (the daemon rewires to an inert delivery path)
rig slack enable Enable the connector (daemon seeds the current backlog as history — no replay storm — then rewires)
rig slack inbound [RETIRED — S10] the in-daemon subsystem owns Socket Mode inbound
rig slack outbound [RETIRED — S10] the in-daemon subsystem owns outbound delivery
rig slack setup Configure the connector (first-class config; secrets stay in the env file, never here)
rig slack status Show the connector's configured + resolvable state (honest; no network)
rig slack verify Live-verify GRANTED Slack scopes (from response headers) + channel membership
rig snapshot Manage rig snapshots
rig snapshot list List snapshots for a rig
rig spec Manage rig specs
rig spec audit Advisory audit of rig-spec culture and startup context
rig spec preflight Run preflight diagnostics on a rig spec
rig spec show Print the spec that recreates a running rig
rig spec validate Validate a rig spec (pure schema validation)
rig specs Browse, preview, and manage the spec library, including managed apps
rig specs add Add a spec file or full spec directory to the user library
rig specs ls List library rigs, agents, and managed apps
rig specs preview Show structured spec review, including managed app details
rig specs remove Remove a user-file spec from the library
rig specs rename Rename a user-file spec in the library
rig specs show Show spec metadata and path
rig specs sync Rescan spec library roots
rig start Start the daemon, verify kernel, and restore rigs that were last running
rig startup-proof Submit and inspect startup orientation proof
rig startup-proof submit Submit this seat's startup proof through the authenticated OpenRig activity hook
rig status Show rig status
rig stream Coordination L1 — append-only intake stream
rig stream archive Soft-archive a stream item (audit row preserved)
rig stream emit Append a stream item
rig stream list List stream items chronologically
rig stream show Fetch one stream item by id
rig stream watch Watch the stream (initial replay + live items)
rig terminal Open OpenRig views (agent terminals) as tiles in a terminal provider (herdr / cmux)
rig terminal open Open every live agent in the view as an interactive terminal tile
rig terminal status Show terminal provider availability + liveness (doctor)
rig terminal views List saved views + the rigs openable as derived views
rig transcript Read agent transcript output
rig tui open mission control (standalone by default; --shared joins the kernel terminal)
rig tui commands list every TUI command (serialized from the ONE command registry; --json for agents)
rig ui UI commands
rig ui open Open the OpenRig UI in the default browser
rig unarchive Unarchive a rig (reverse of 'rig archive'): returns it to the default view. Always safe.
rig unclaim Release an adopted session without killing the tmux session
rig up Launch a rig or managed app from a spec, library entry, or bundle
rig usage Per-seat token telemetry over time (series + top-N burn) — facts for the oversight detector
rig usage series raw per-seat usage samples, oldest first
rig usage top top-N seats by token burn over the window
rig view Coordination L5 — daemon-backed views over coordination state
rig view list List built-in + custom views
rig view register Register or update a custom view
rig view show Run a view (built-in or custom). Built-in views: recently-active, founder, pod-load, escalations, held, activity, pickup
rig walk Walk a seat through a paced sequence of context pieces
rig watchdog Coordination Watchdog — daemon-native scheduler for reminders, artifact gates, workflow health, idle gates, and context usage
rig watchdog list List watchdog jobs (default: active + compact + at most 100)
rig watchdog register Register a watchdog; queue block --wake-watchdog attaches its job id. Context transcripts measured 113K–153K tokens/MB. The margin is the protection because prompt-bound consumers act only at turn boundaries
rig watchdog show Show one watchdog job
rig watchdog status Show one watchdog job + recent evaluation history
rig watchdog stop Stop a watchdog job (operator-stopped; scheduler skips it)
rig whoami Show current managed identity in an OpenRig topology
rig workflow Daemon-native Workflow Runtime — declarative spec + transactional-scribe step projection (PL-004 Phase D)
rig workflow abort Transactionally cancel every live packet and abort the whole workflow instance
rig workflow compile Compile project.yaml → mission.yaml → slice.yaml into an inspectable lifecycle graph (read-only)
rig workflow continue Inspect an instance's current frontier + step trail (read-only; advancing happens via 'rig workflow project')
rig workflow guidance Read current selected SDLC teaching, original intent and authored/bound provenance
rig workflow instantiate Create a workflow instance + entry-step qitem from a spec
rig workflow instantiate-lifecycle Compile and instantiate an eligible lifecycle; recover a lost response with workflow operation <key>
rig workflow list List workflow instances; optionally filter by status
rig workflow operation Recover a lifecycle creation or revision effect by its stable key, even after a lost response or source edit
rig workflow project Close a current packet AND project the next-step packet (transactional-scribe; one daemon transaction)
rig workflow resume Redrive a FAILED instance from its failed step (completed steps never re-run; one fresh max_hops window)
rig workflow revise Compare authored and running lifecycle graphs; deliberately adopt compatible changes without replaying work
rig workflow route Re-route the current frontier step to a new owner (same step, honest handoff closure; never advances)
rig workflow run Instantiate a workflow AND follow it live to a terminal state (exit 0 completed / 3 failed)
rig workflow show Show current work, exception-owner readiness and existing exception obligations
rig workflow specs List registered workflow specs; built-in starters tagged with (built-in)
rig workflow status Which instances need attention: counts + one row per failed/stuck/waiting instance with reason + next action (read-only)
rig workflow trace Show one workflow instance + its append-only step trail (audit-only verdict)
rig workflow validate Validate a workflow spec file (returns structured ok/error report)
rig workflow watch Attach to an in-flight instance and follow it live (read-only; exit mirrors the outcome)
rig workspace PL-007 Workspace Primitive — typed-kind tooling. validate walks a root and reports frontmatter gaps; doctor runs the 8-check workspace-readiness diagnostic.
rig workspace doctor Run the 8-check workspace-readiness diagnostic against the daemon's resolved workspace. Reports state of workspace root, missions folder, file allowlist, daemon alignment, daemon reload, slice docs, mission NOTES, and SDLC convention sections. Read-only.
rig workspace validate Walk a workspace root, parse each .md file's YAML frontmatter, and emit a structured gap report. Advisory only — never modifies files. Default root: cwd.

Global options

Flag Purpose / help default
-h, --help display help for command
-V, --version output the version number

rig add

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--rig-root <path> Root directory for agent resolution

rig adopt

Flag Purpose / help default
--bind <logicalId=tmuxSessionOrDiscoveryId> Bind a logical node to a discovered tmux session or discovery ID (default: [])
--bindings-file <path> Load logicalId -> tmux session/discovery ID mappings from YAML
-h, --help display help for command
--json Output machine-readable JSON
--rig-root <root> Root directory for pod-aware resolution
--target-rig <rigId> Target existing rig for additive materialization

rig agent

Flag Purpose / help default
-h, --help display help for command

rig agent-image

Flag Purpose / help default
-h, --help display help for command

rig agent-image create

Flag Purpose / help default
--estimated-tokens <n> Operator-supplied token estimate
-h, --help display help for command
--image-version <version> Image version (default: 1)
--json JSON output
--lineage <names...> Comma-separated parent image names if forking from another image
--name <name> Image name (used as the directory name and library id)
--notes <text> Operator-supplied notes preserved in manifest

rig agent-image delete

Flag Purpose / help default
--force Override evidence-preservation guard (CATASTROPHIC if active references exist)
-h, --help display help for command
--json JSON output

rig agent-image list

Flag Purpose / help default
-h, --help display help for command
--json JSON output
--runtime <runtime> Filter by runtime (claude-code | codex)

rig agent-image pin

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig agent-image preview

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig agent-image prune

Flag Purpose / help default
--dry-run Preview without deleting (default if not --force)
--force Override evidence-preservation guard (CATASTROPHIC: deletes referenced images)
-h, --help display help for command
--json JSON output

rig agent-image show

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig agent-image sync

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig agent-image unpin

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig agent validate

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig archive

Flag Purpose / help default
--force Archive even if the rig is running or degraded
-h, --help display help for command
--json JSON output for agents

rig ask

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--runtime <runtime> runtime for --wake: claude (default) or codex
--seat <session-name> Scope the search to ONE seat's transcript across every generation that sat in it (cross-generation archaeology)
--session <token> Search ONE specific session's JSONL by its session token (read-only)
--wake <seat[@gen]|token> EXECUTES: wake a session (by seat[@generation] or raw resume token) — ask one question, get a snapshot answer, back to cold (runtime cost; explicit, never an implicit escalation from a search)
--wake-timeout <seconds> bounded wake timeout in seconds (default 180)

rig attach

Flag Purpose / help default
--cwd <path> Working directory to record (default: current working directory)
--display-name <name> External session/display name to record
-h, --help display help for command
--json JSON output
--member <name> Member name for pod attach mode
--node <logicalId> Attach to an existing logical node
--pod <namespace> Attach by creating a new member in an existing pod
--print-env Print shell exports for OPENRIG_NODE_ID and OPENRIG_SESSION_NAME
--rig <rigId> Target rig ID
--runtime <runtime> Runtime for pod attach mode; optional guard for node mode
--self Attach the current shell/agent (required in v1)

rig auth

Flag Purpose / help default
-h, --help display help for command

rig auth list

Flag Purpose / help default
-h, --help display help for command
--runtime <runtime> Runtime axis (MVP: codex) (default: "codex")

rig auth save

Flag Purpose / help default
-h, --help display help for command
--runtime <runtime> Runtime axis (MVP: codex) (default: "codex")

rig auth seats

Flag Purpose / help default
-h, --help display help for command

rig auth seats list

Flag Purpose / help default
-h, --help display help for command
--runtime <runtime> Runtime axis (MVP: codex) (default: "codex")

rig auth seats report

Flag Purpose / help default
-h, --help display help for command
--runtime <runtime> Runtime axis (MVP: codex) (default: "codex")

rig auth seats set

Flag Purpose / help default
--cwd <cwd> Working dir (metadata)
-h, --help display help for command
--profile <profile> Auth profile label (or omit for unknown)
--rig <rig> Rig name
--runtime <runtime> Runtime axis (MVP: codex) (default: "codex")
--seat <seat> Seat session name

rig auth seats show

Flag Purpose / help default
-h, --help display help for command
--runtime <runtime> Runtime axis (MVP: codex) (default: "codex")

rig auth status

Flag Purpose / help default
-h, --help display help for command
--runtime <runtime> Runtime axis (MVP: codex) (default: "codex")

rig auth switch

Flag Purpose / help default
-h, --help display help for command
--runtime <runtime> Runtime axis (MVP: codex) (default: "codex")

rig auth validate

Flag Purpose / help default
-h, --help display help for command
--runtime <runtime> Runtime axis (MVP: codex) (default: "codex")

rig bind

Flag Purpose / help default
-h, --help display help for command
--member <name> Member name for the new node (requires --pod)
--node <logicalId> Bind to existing logical node
--pod <namespace> Create new node in this pod (requires --member)
--rig <rigId> Target rig ID

rig bootstrap

Flag Purpose / help default
--cwd <path> Override launch working directory for all members for this run only
-h, --help display help for command
--json Output as parseable JSON
--plan Plan mode — show reviewed plan without executing
--yes Auto-approve trusted deterministic actions

rig broadcast

Flag Purpose / help default
--context <ref> Broadcast a context pack by its path-like ref (e.g. packs/fleet-update). Resolved content is fanned out; an oversized ref is flagged 'walk-sized'.
--force Send even if targets appear mid-task
-h, --help display help for command
--host <id> Broadcast on a remote host declared in ~/.openrig/hosts.yaml (http hosts only — CLI-direct to the remote daemon's fan-out)
--json JSON output for agents
--pod <name> Broadcast to all sessions in a pod
--rig <name> Broadcast to all sessions in a rig

rig bundle

Flag Purpose / help default
-h, --help display help for command

rig bundle create

Flag Purpose / help default
--allow-drift Bundle a spec that disagrees with the running rig of the same name; the divergence is stamped into bundle provenance
--bundle-version <ver> Bundle version (default: "0.1.0")
-h, --help display help for command
--include-packages <refs...> Package refs to include (default: all from spec)
--json JSON output
--min-cli-version <ver> Minimum CLI version required to install this bundle (Item 2 compatibility)
--min-daemon-version <ver> Minimum daemon version required to install this bundle (Item 2 compatibility)
--name <name> Bundle name (default: "my-bundle")
--notes <text> Operator notes captured in bundle provenance metadata
-o, --output <path> Output path for .rigbundle
--rig-root <root> Root directory for pod-aware resolution

rig bundle history

Flag Purpose / help default
-h, --help display help for command
--json JSON output
--rig <name> Filter to records whose targetRigName matches
--since <iso> Filter to records installedAt >= this ISO timestamp

rig bundle inspect

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig bundle install

Flag Purpose / help default
--force Operator-explicit override of the Item-3 install-time conflict check (NOT recommended; conflicts may produce partial install state)
-h, --help display help for command
--json JSON output
--plan Plan mode
--skip-version-check Operator-explicit override of the Item-2 install-time compatibility check (NOT recommended for routine use)
--target <root> Target root directory
--yes Auto-approve

rig capture

Flag Purpose / help default
-h, --help display help for command
--host <id> Capture on a remote host declared in ~/.openrig/hosts.yaml (ssh hosts shell out; http hosts go CLI-direct to the remote daemon)
--json JSON output for agents
--lines <n> Number of lines to capture (default: 20) (default: "20")
--pod <name> Capture all sessions in a pod
--rig <name> Capture all sessions in a rig

rig chatroom

Flag Purpose / help default
-h, --help display help for command

rig chatroom clear

Flag Purpose / help default
-h, --help display help for command

rig chatroom history

Flag Purpose / help default
--after <id> Messages after this message ID
-h, --help display help for command
--json JSON output
--limit <n> Limit results (default: "50")
--sender <name> Messages from this sender
--since <timestamp> Messages since this timestamp
--topic <name> Filter by topic

rig chatroom send

Flag Purpose / help default
-h, --help display help for command
--sender <name> (deprecated, ignored) the sender is derived from the seat env (X-OpenRig-Session); P21 made the chat route derive it from the transport header

rig chatroom topic

Flag Purpose / help default
--body <text> Optional body text
-h, --help display help for command
--sender <name> (deprecated, ignored) the sender is derived from the seat env (X-OpenRig-Session); the chat topic route derives it from the transport header

rig chatroom wait

Flag Purpose / help default
--after <id> Only messages after this ID
-h, --help display help for command
--json JSON output
--sender <name> Filter by sender
--timeout <seconds> Timeout in seconds (default: "120")
--topic <name> Filter by topic

rig chatroom watch

Flag Purpose / help default
-h, --help display help for command
--tmux Run watch in a dedicated tmux session

rig compact

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig compact-plan

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--refresh Re-sample context usage before planning
--rig <name> Plan one rig only
--threshold-percent <0-100> Used-percent threshold when context window size is missing
--threshold-tokens <n> Estimated used-token threshold for Claude compact-plan candidates

rig config

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents (resolved RiggedConfig)
--with-source Include source/default per key (honest provenance)

rig config get

Flag Purpose / help default
-h, --help display help for command
--json JSON output with value + source + default
--show-source Print value + source on a single line

rig config init-workspace

Flag Purpose / help default
--dry-run Show what would be created without writing
--force Deprecated compatibility flag; existing files are still preserved
-h, --help display help for command
--json JSON output
--root <path> Override workspace root (default: workspace.root setting)

rig config reset

Flag Purpose / help default
-h, --help display help for command

rig config set

Flag Purpose / help default
-h, --help display help for command

rig context

Flag Purpose / help default
-h, --help display help for command

rig context add

Flag Purpose / help default
--checkout With --git, select an existing checkout instead of cloning; updates may merge in it
--git Clone a Git repository path/URL with existing Git credentials; select a pack snapshot
-h, --help display help for command
--json JSON output
--name <name> Override the install name (defaults to the manifest name / source basename)
--pack <path> With --git, select a repository-relative pack; default discovers manifest.yaml or .openrig/context-packs

rig context compose

Flag Purpose / help default
--from <files...> Ordered source files
-h, --help display help for command
--out <ref> Path-like durable output ref

rig context get

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig context list

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig context preview

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig context profile

Flag Purpose / help default
--budget <tokens> Situation token budget — overage is REPORTED, never truncated
-h, --help display help for command
--json JSON output (the full composed profile)
--mission <mission> Grant read access to that mission's tree (mission: atoms)
--profile <profile> Named install profile declared by the pack (selection + ordered phases)
--rig <rig> With --seat: grant read access to that seat's tree (seat: atoms)
--runtime <runtime> claude-code (alias: claude) or codex (default: $OPENRIG_RUNTIME, else claude-code)
--seat <seat> With --rig: the seat whose tree seat: atoms may read
--situation <situation> fresh | handover | post-compaction
--slice <slice> With --mission: compose the legacy default project/mission/slice SPEC walk

rig context recap-write

Flag Purpose / help default
--file <path> Markdown file containing the recap content
-h, --help display help for command
--rig <rig> Rig name (the rigs/<rig> altitude)
--seat <seat> Seat id (the seats/<seat> altitude)

rig context rm

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig context show

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig context source

Flag Purpose / help default
-h, --help display help for command

rig context source inspect

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig context source update

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig context sync

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig context trace

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--name <file> Chain filename, identical at every altitude (e.g. LEARNED.md, CULTURE.md)
--pod <pod> Pod id (the pods/<pod> altitude); omit when no pod context is selected
--rig <rig> Rig name (the rigs/<rig> altitude)
--seat <seat> Seat id (the seats/<seat> altitude); omit for a rig-level trace

rig context work-install

Flag Purpose / help default
--apply-skills Reconcile selected skills into the runtime harness directory
--cwd <path> Agent working directory that receives skill projections (default: current directory)
--deliver Include the exact content of each extant planned file
-h, --help display help for command
--json JSON output
--mission <id> Exact mission id under the selected project
--project <id> Exact project id from workspace.yaml
--runtime <runtime> Inspect skills for claude-code (alias: claude) or codex
--slice <id> Exact slice id under the selected mission
--topology <ids> Comma-separated topology/profile skill identities

rig crash-cart

Flag Purpose / help default
-h, --help display help for command
--json emit JSON (the default machine-readable form)

rig create

Flag Purpose / help default
--cwd <path> Working directory for the seat (default: current working directory)
-h, --help display help for command
--json JSON output for agents
--runtime <runtime> Agent runtime (default: "claude-code")

rig daemon

Flag Purpose / help default
-h, --help display help for command

rig daemon logs

Flag Purpose / help default
--follow Follow log output
-h, --help display help for command

rig daemon start

Flag Purpose / help default
--db <path> Database path
-h, --help display help for command
--host <host> Host to bind on
--no-kernel Skip kernel auto-boot (daemon serves without the kernel rig)
--port <port> Port to listen on
--wait-for-kernel After daemon binds, also wait for kernel-agent readiness (default timeout 60s)
--wait-for-kernel-ms <ms> Override --wait-for-kernel timeout in milliseconds

rig daemon status

Flag Purpose / help default
-h, --help display help for command

rig daemon stop

Flag Purpose / help default
-h, --help display help for command

rig destroy

Flag Purpose / help default
--all Destroy OpenRig state and remove managed tmux sessions
--backup Move state aside instead of deleting it
--confirm <token> Exact confirmation token: destroy-openrig-state
-h, --help display help for command
--state Destroy OpenRig state and recreate an empty state root
--yes Confirm the destructive operation

rig discover

Flag Purpose / help default
--draft Generate a candidate rig spec from discovered sessions
-h, --help display help for command
--json Output as parseable JSON

rig doctor

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--spec <path> Rig spec to check against the running rig of the same name (spec-vs-live topology)

rig down

Flag Purpose / help default
--delete Delete rig record after stopping
--force Kill sessions immediately
-h, --help display help for command
--host <id> Run on a remote host declared in ~/.openrig/hosts.yaml
--json JSON output for agents
--snapshot Take snapshot before teardown

rig env

Flag Purpose / help default
-h, --help display help for command

rig env down

Flag Purpose / help default
-h, --help display help for command
--volumes Also remove volumes

rig env logs

Flag Purpose / help default
-h, --help display help for command
--tail <n> Number of lines (default: "100")

rig env status

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig expand

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--rig-root <path> Root directory for agent resolution

rig export

Flag Purpose / help default
-h, --help display help for command
-o, --output <path> Output file path (default: "rig.yaml")

rig file

Flag Purpose / help default
-h, --help display help for command

rig file copy

Flag Purpose / help default
--dry-run Print the exact planned transfer (src, dst, host, files/bytes) and move nothing
-h, --help display help for command
--json JSON output for agents

rig fork

Flag Purpose / help default
-h, --help display help for command
--image-name <name> Name for the kept image (with --keep-image; default fork-<member>)
--image-version <version> Version for the kept image (default 1)
--json JSON output for agents
--keep-image Also capture a durable, pinned agent image (default forks one-shot with no library growth)
--member <member-id> New member id for the forked successor
--pod <pod-namespace> Target existing pod namespace
--rig <rig-id> Target rig id
--rig-root <path> Root directory for agent resolution

rig gateway

Flag Purpose / help default
-h, --help display help for command

rig gateway human

Flag Purpose / help default
-h, --help display help for command

rig gateway human add

Flag Purpose / help default
--actor <actor> Named operator when outside a managed seat
--away Set the AWAY preset
--binding <kind:connectorRef:secretsRef:role[:handle=<id>]> Connector binding (repeatable); role primary|secondary, EXACTLY ONE primary. Optional handle=<platform id> makes the binding inbound-resolvable (unique per kind); omit it for outbound-only.
--delivery-class <A|B|C|D> Notification loudness class (the notifications register selection)
--display-name <name> Human-readable display name
-h, --help display help for command
--reason <reason> Reason recorded with connector binding changes (default: "register human delivery")
--replace Explicitly replace an existing human (no silent overwrite)

rig gateway human list

Flag Purpose / help default
-h, --help display help for command
--json Complete record(s) as JSON

rig gateway human remove

Flag Purpose / help default
--force Archive despite KNOWN in-flight items (each is recorded as orphaned)
-h, --help display help for command

rig gateway human set

Flag Purpose / help default
--actor <actor> Named operator when outside a managed seat
-h, --help display help for command
--reason <reason> Reason recorded with connector binding changes (default: "update human binding")

rig gateway human show

Flag Purpose / help default
-h, --help display help for command
--json Complete record as JSON

rig grow

Flag Purpose / help default
--cwd <path> Working directory for the seat (default: current working directory)
-h, --help display help for command
--json JSON output for agents
--new-pod <pod> Create a new pod for the seats
--pod <pod> Target pod; inferred when the rig has one pod
--runtime <runtime> Agent runtime (default: "claude-code")

rig handover

Flag Purpose / help default
--dry-run Plan the handover without changing topology
-h, --help display help for command
--json JSON output for agents
--operator <address> Operator initiating the handover
--reason <reason> Why the handover is happening
--source <source> Successor source: fresh (default; launches a new agent), discovered:<id> (operator-prepared), fork:<id> (native fork of the source conversation), or rebuild (fresh agent primed from the seat's durable artifacts).

rig health

Flag Purpose / help default
--actor <name> Attribute explicit diagnosis writes when operating outside a managed seat
-h, --help display help for command
--instance [instance-id] Inspect the whole local instance, or one canonical instance scope by ID
--json JSON output; diagnosis list/show need --full for complete evidence
--limit <count> Maximum findings (1-200) (default: 100)
--rig <rig-id> Inspect findings whose canonical scope is this rig
--seat <node-id> Inspect one seat by stable node ID
--self Inspect the current seat (default)
--severity <severity> Filter by severity (choices: "info", "warning", "critical")
--status <status> Filter by finding status (choices: "active", "cleared", "indeterminate")

rig health checkpoint

Flag Purpose / help default
--file <path> Submit checkpoint JSON with exact queue and product evidence
-h, --help display help for command
--json No description supplied by help.

rig health diagnose

Flag Purpose / help default
--apply No description supplied by help.
-h, --help display help for command
--json No description supplied by help.

rig health diagnosis

Flag Purpose / help default
-h, --help display help for command

rig health diagnosis list

Flag Purpose / help default
--full Complete records including all evidence and receipts; may be large
-h, --help display help for command
--json Summary array with explicit omitted fields and per-occurrence expansion commands

rig health diagnosis notify

Flag Purpose / help default
-h, --help display help for command
--json No description supplied by help.

rig health diagnosis record

Flag Purpose / help default
--file <path> Disposition JSON with verdict, causalStart, steering, uncertainty, evidenceRefs
-h, --help display help for command
--json No description supplied by help.

rig health diagnosis show

Flag Purpose / help default
--full Complete original record; use --full --json for lossless JSON (may be large)
-h, --help display help for command
--json Summary JSON with omitted fields, original byte size and exact full command

rig health explain

Flag Purpose / help default
-h, --help display help for command
--json Emit the canonical daemon health record as JSON

rig health policy

Flag Purpose / help default
--file <path> Apply policy JSON with an audit record
-h, --help display help for command
--json No description supplied by help.

rig heartbeat

Flag Purpose / help default
-h, --help display help for command
--include-done Include done/handed-off queue items in output
--json JSON output for agents
--nudge Send informational nudges to stalled/unproven owners
--rig <name> Limit to a single rig

rig host

Flag Purpose / help default
-h, --help display help for command

rig host add

Flag Purpose / help default
--bearer-env <name> Env var NAME holding the bearer token — http transport (optional pointer, never a value; omit both bearer flags for a tokenless daemon)
--bearer-file <path> File PATH holding the bearer token — http transport (optional pointer, never a value; at most one bearer flag)
-h, --help display help for command
--id <id> Unique host id
--json JSON output
--notes <text> Free-form operator note
--target <target> SSH target (DNS name, ssh-config alias, or IP) — ssh transport
--transport <transport> ssh or http
--url <url> Remote daemon base URL — http transport
--user <user> SSH user — ssh transport

rig host doctor

Flag Purpose / help default
-h, --help display help for command
--json JSON output
--posture <profile> Run the posture baseline (only built-in profile: product-factory-vps)
--public-addr <ip> Public address for the outside-vantage reachability probes (posture items 8-9)

rig host list

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig host pair

Flag Purpose / help default
-h, --help display help for command
--human <address> Registered @external approval recipient on the target (required when several exist)
--id <id> Registry id for the new host (default: derived from the hostname)
--json JSON output
--timeout <seconds> How long to wait for the target-side approval (default: "600")

rig host rename

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig host select

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig import

Flag Purpose / help default
--cwd <path> Override launch/materialization working directory for all members
-h, --help display help for command
--instantiate Instantiate the rig after import
--materialize-only Create rig topology without launching sessions
--preflight Run preflight checks
--rig-root <root> Root directory for pod-aware resolution
--target-rig <rigId> Target existing rig for materialization or workspace apply
--workspace-only Apply only the workspace declaration to an existing rig

rig launch

Flag Purpose / help default
-h, --help display help for command
--hold-reason <reason> Reason for holding non-target seats
--host <id> Run on a remote host declared in ~/.openrig/hosts.yaml
--json JSON output
--plan Show subset selection and non-target effects without mutation
--seats <ids> Comma-separated logical IDs for subset launch
--snapshot-id <id> Use this exact restore-usable snapshot

rig mcp

Flag Purpose / help default
-h, --help display help for command

rig mcp serve

Flag Purpose / help default
-h, --help display help for command
--port <port> Daemon port override

rig mode

Flag Purpose / help default
-h, --help display help for command

rig mode cite

Flag Purpose / help default
-h, --help display help for command
--mission <id> No description supplied by help.
--project <id> No description supplied by help.
--qitem <id> No description supplied by help.
--rig <id> No description supplied by help.
--workstream <id> No description supplied by help.

rig mode defaults

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig mode effective

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--mission <id> No description supplied by help.
--project <id> No description supplied by help.
--qitem <id> No description supplied by help.
--rig <id> No description supplied by help.
--workstream <id> No description supplied by help.

rig mode set

Flag Purpose / help default
--autonomy-scope <v> No description supplied by help.
--bearer <token> Operator bearer token (or set OPENRIG_AUTH_BEARER_TOKEN env).
--concurrency-limit <v> No description supplied by help.
--confirm Confirm the proposed binding and apply it. Without this flag, set is restate-only.
--escalation-threshold <v> No description supplied by help.
--evidence <citation> Free-text citation (operator message id, file pointer, chatroom topic, etc.).
--expiry-or-stale-rule <v> No description supplied by help.
--heartbeat-cadence <v> No description supplied by help.
-h, --help display help for command
--inspection-depth <v> No description supplied by help.
--json JSON output for agents
--permission-prompt-posture <v> One of: normal | batch_for_human | do_not_prompt_unless_blocked (auto_accept is FORBIDDEN by convention).
--qualifier <id> Rig/project/qitem ID; mission: project/mission; workstream: project/mission/slice-id; omit for global_host
--scope <scope> Scope: global_host | rig | project | mission | workstream | qitem (default: per-mode recommendation)
--update-detail <v> No description supplied by help.

rig mode show

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig mode unset

Flag Purpose / help default
--bearer <token> No description supplied by help.
-h, --help display help for command
--json JSON output for agents

rig package

Flag Purpose / help default
-h, --help display help for command

rig package install

Flag Purpose / help default
--allow-merge Allow managed block merges into existing files
-h, --help display help for command
--role <name> Role to install
--runtime <runtime> Runtime (claude-code or codex) (default: "claude-code")
--target <dir> Target repository root (default: ".")

rig package list

Flag Purpose / help default
-h, --help display help for command

rig package plan

Flag Purpose / help default
-h, --help display help for command
--role <name> Role to install
--runtime <runtime> Runtime (claude-code or codex) (default: "claude-code")
--target <dir> Target repository root (default: ".")

rig package rollback

Flag Purpose / help default
-h, --help display help for command

rig package validate

Flag Purpose / help default
-h, --help display help for command

rig parked

Flag Purpose / help default
-h, --help display help for command
--json Full diagnosis as JSON
--rig <rig> Rig scope (defaults to your seat's rig from OPENRIG_SESSION_NAME; a seat argument carrying @rig self-scopes)

rig plugin

Flag Purpose / help default
-h, --help display help for command

rig plugin list

Flag Purpose / help default
-h, --help display help for command
--json JSON output
--runtime <runtime> Filter by runtime support: claude | codex (omit for all)
--source <source> Filter by source: vendored | claude-cache | codex-cache

rig plugin show

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig plugin used-by

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig plugin validate

Flag Purpose / help default
-h, --help display help for command
--json JSON output ({ valid: boolean, errors: string[] })

rig policy

Flag Purpose / help default
-h, --help display help for command

rig policy apply

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output
--spec <path> Rig spec file, or a directory containing rig.yaml/agent.yaml

rig policy current

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output
--spec <path> Rig spec file, or a directory containing rig.yaml/agent.yaml

rig policy list

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output
--spec <path> Rig spec (file or directory) defining the custom-policy context

rig policy show

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output
--spec <path> Rig spec (file or directory) whose directory anchors a custom ref

rig preflight

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig project

Flag Purpose / help default
-h, --help display help for command

rig project classify

Flag Purpose / help default
--action <action> Action type (e.g., create, advance)
--confidence <confidence> Classification confidence (e.g., low, medium, high)
--destination <destination> Classification destination (downstream slice/seat)
-h, --help display help for command
--json JSON output for agents
--maturity <maturity> Classification maturity (e.g., concept, drafted, ratified)
--session <session> Classifier session name (must hold active lease)
--type <type> Classification type (e.g., idea, bug, feature-request)
--urgency <urgency> Classification urgency (e.g., normal, high, critical)

rig project lease-acquire

Flag Purpose / help default
--evaluate-deadness-first Before acquire, call evaluateDeadness to clear any stale TTL-expired or dead-holder lease (per PRD § L2 deadness-detection)
-h, --help display help for command
--json JSON output for agents
--session <session> Classifier session name

rig project lease-heartbeat

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--lease-id <id> Lease ID
--session <session> Classifier session name

rig project lease-show

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig project list

Flag Purpose / help default
--destination <destination> Filter by classification destination
-h, --help display help for command
--json JSON output for agents
--limit <n> Result limit (default: "100")
--session <session> Filter by classifier session

rig project reclaim-classifier

Flag Purpose / help default
-h, --help display help for command
--if-dead Only reclaim if the current holder is reported dead by the liveness check
--json JSON output for agents
--reason <text> Reclaim reason (free-form; recorded in classifier_leases.reclaim_reason)
--session <session> Session that will hold the new lease

rig project show

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig proof

Flag Purpose / help default
-h, --help display help for command
--workspace <path> Capture/add only: override workspace root; show/judge use the selected daemon workspace

rig proof add

Flag Purpose / help default
--artifact-type <type> C1 artifact_type, one of: guard | qa | rev1-r1 | rev1-r2 | adjudication
--body <text> Artifact body inline (mutually exclusive with --file)
--candidate-sha <sha> C1 candidate_sha — the join key (convention C2): the proven candidate tip this artifact judges
--evidences <refs> D2 attestation: comma-separated proof-contract item refs this artifact covers (item text or 1-based index)
--file <path> Artifact body from a file (mutually exclusive with --body)
-h, --help display help for command
--json JSON output for agents
--media <refs> Corrective §3.4: comma-separated media refs (relative to the slice proof/ dir) this drop stands behind — appended to the artifact body as markdown refs so the composer curates them into delivered.items[].proof
--mission <name> Hint mission when slice-path is just NN-slug
--money-evidence <line> C1 money_evidence — the one line of money evidence
--name <filename> Artifact filename in proof/ (defaults to the --file basename, else <artifact-type>-<verdict>-<UTC>.md)
--self-check <text> D2 attestation: the agent's assertion that it LOOKED at the evidence and confirmed it shows the claim
--slice-id <dot-id> C1 slice dot-ID (defaults to the slice frontmatter id)
--verdict <verdict> C1 verdict, one of: CLEAR | BLOCKING | CONCERNING | PASS | NOT-CLEAR

rig proof judge

Flag Purpose / help default
--comparison <ref> Patch-equivalent subject: actual comparison/adoption receipt alongside outcome evidence
--evidence <ref> Existing evidence, relative to slice or workspace missions/; repeat for multiple references (default: [])
-h, --help display help for command
--json Return committed receipt and current readiness
--operation-id <id> Advanced: explicit retry identity
--reason <text> The evidence-backed judgment being made
--replace Deliberately reaffirm a historical judgment after a later correction
--revision <revision> Advanced: deliberately require this item revision
--subject <kind:ref> artifact, commit or patch-equivalent subject; artifact inferred when omitted
--verdict <verdict> accept | reject | withdraw

rig proof show

Flag Purpose / help default
-h, --help display help for command
--json Structured readiness, item revisions and retained judgment references

rig provider

Flag Purpose / help default
-h, --help display help for command

rig provider accounts

Flag Purpose / help default
--account <a> Filter by account ref
-h, --help display help for command
--json Output as parseable JSON
--provider <p> Filter by provider (codex|claude)

rig provider bindings

Flag Purpose / help default
--account <a> Filter by account ref
-h, --help display help for command
--json Output as parseable JSON
--provider <p> Filter by provider (codex|claude)

rig provider precheck

Flag Purpose / help default
-h, --help display help for command
--json Output as parseable JSON
--seat <s> The seat session
--to-account <a> The target account ref

rig provider signals

Flag Purpose / help default
--account <a> Filter by account ref
-h, --help display help for command
--json Output as parseable JSON
--provider <p> Filter by provider (codex|claude)

rig provider status

Flag Purpose / help default
-h, --help display help for command
--json Output as parseable JSON

rig provider switch

Flag Purpose / help default
--force-unsafe Override a non-stranding precheck failure (still refuses to strand a live conversation)
-h, --help display help for command
--json Output as parseable JSON
--seat <s> The seat session
--to-account <a> The target account ref

rig ps

Flag Purpose / help default
--active Shortcut for --filter agentActivity.state=running (PL-019)
--all-hosts Fan out to all registered HTTP hosts (observation-only)
-A, --all-rigs Show all rigs (default is current rig only)
--fields <list> Comma-separated field list to project (JSON only)
--filter <key=value> Filter entries; supported keys: status, lifecycleState, name-prefix, name, agentActivity.state
--full Show all node-list fields per node (uncompacted rows; node-list recoveryGuidance/currentUsage live on the node detail, not the list)
-h, --help display help for command
--host <id> Run on a remote host declared in ~/.openrig/hosts.yaml
--hosts <ids> Fan out to specific hosts (comma-separated)
--include-archived Include archived rigs (default hides them); parity with 'rig stream list --include-archived'
--json JSON output for agents
--limit <n> Limit number of entries (rigs or nodes)
--nodes Show per-node detail (current rig; -A for all rigs)
--rig <name> Show only nodes belonging to the named rig
--running Alias for --active
--session <name> Show only the node matching this canonical session name
--summary Emit aggregate-only output (counts by status/lifecycle)
--verbose Alias for --full

rig queue

Flag Purpose / help default
-h, --help display help for command

rig queue block

Flag Purpose / help default
--actor <session> (deprecated, ignored) the actor is derived from the seat env (X-OpenRig-Session); the park writes via the same P21 I3 header-deriving update route
--continuation <text> What resumes. Workspace deferred/not-imminent work belongs in a mission/slice
--evidence-ref <path> Durable artifact the human judges (required for human-seat parks unless already on the item)
-h, --help display help for command
--json JSON output for agents
--note <text> Transition note for the audit log
--on <blocker> The blocker: a live blocker qitem, typed gate, or human-seat session
--summary <text> Plain-language summary of the decision owed (required for human-seat parks unless already on the item)
--wake-after <duration> Atomically arm a timer with the park (for example 90s, 15m, 2h)
--wake-watchdog <jobId> Attach an existing live watchdog id targeting the parked owner

rig queue claim

Flag Purpose / help default
--destination <session> (deprecated, ignored) the claimant is derived from the seat env (X-OpenRig-Session); P21 I3 made the claim route derive it from the transport header
-h, --help display help for command
--json JSON output for agents

rig queue create

Flag Purpose / help default
--body <text> Qitem body inline (use - to read from stdin; mutually exclusive with --body-file)
--body-context <ref> Snapshot a context pack by its path-like ref into the qitem body (the resolved content rides the handoff + a body-context:<ref> provenance tag). Mutually exclusive with --body / --body-file.
--body-file <path> Read qitem body from a file path (use - for stdin; mutually exclusive with --body). Kills the backtick-shell-corruption class for multiline bodies.
--destination <session> Destination session (the seat that owns the work)
--evidence-ref <path> OPR.0.4.4.19 FR-5: pointer to the durable artifact a human judges (e.g. a PROOF.md path). Required by the daemon when the item is human-routed; optional otherwise.
--expires-at <iso> ISO timestamp at which the qitem expires
--gate <role> OPR.0.4.3.16: mark this as a gate qitem; translated to a gate:<role> tag (role e.g. guard | spec-review | pm-lead | qa | human). The idle-gate watchdog reads this predicate. Composes with --tags.
-h, --help display help for command
--host <id> OPR.0.4.6.MH3: route this queue write to a REGISTERED remote host (see rig host ls). EXPLICIT-ONLY — queue verbs never follow the persisted 'rig host select' selection. Equivalent to the host-qualified destination form example-seat@example@<host>.
--human-detail-file <path> One explicitly authored supplemental thread reply; keep the complete action/options in --body-file
--human-intent <intent> decision (default) or update: a quiet informational delivery, never an approval request
--id <qitemId> Idempotent qitem_id (skip if not provided)
--json JSON output for agents
--mission <id> First-class mission scope; translated to a mission:<id> tag (composes with --tags)
--no-nudge Suppress the default destination nudge (cold-queue)
--priority <priority> Priority: routine | urgent | critical (default: "routine")
--slice <id> First-class slice scope; translated to a slice:<id> tag (composes with --tags)
--source <session> (deprecated, ignored) the source is derived from the seat env (X-OpenRig-Session); P21 I3 made the create route derive it from the transport header
--summary <text> Short human-readable subject, shown in the needs-you view. For a human destination, --body-file is the complete decision brief or update; keep technical continuation in the owning agent row and evidence.
--tags <tags> Comma-separated tags (composes with --mission and --slice)
--target-repo <name> PL-007: typed repo scope (must match a repo in the source rig's RigSpec.workspace.repos[])
--tier <tier> Tier (e.g. fast, routine, deep, critical) — drives SLA
--verify Boundedly wait for the existing gateway delivery receipt after persistence; never retries the create and never claims human readership

rig queue fallback

Flag Purpose / help default
--destination <session> Fallback destination seat
-h, --help display help for command
--json JSON output for agents
--reason <text> Reason for fallback (default: "manual")

rig queue handoff

Flag Purpose / help default
--body <text> New qitem body inline (use - to read from stdin; mutually exclusive with --body-file). Omit both to keep the source body.
--body-file <path> Read the new qitem body from a file path (use - for stdin; mutually exclusive with --body). Kills the backtick-shell-corruption class.
--evidence-ref <path> OPR.0.4.4.19 FR-5: durable-artifact pointer for the new qitem. Required by the daemon when the new qitem is human-routed; optional otherwise.
--from <session> (deprecated, ignored) the handing-off seat is derived from the seat env (X-OpenRig-Session); P21 I3 made the handoff route derive it from the transport header
--gate <role> OPR.0.4.3.16: mark the new qitem as gate work; translated to a gate:<role> tag (e.g. guard | spec-review). The idle-gate watchdog reads this predicate. Composes with --tags.
-h, --help display help for command
--host <id> OPR.0.4.6.MH3: route this queue write to a REGISTERED remote host (see rig host ls). EXPLICIT-ONLY — queue verbs never follow the persisted 'rig host select' selection. Equivalent to the host-qualified destination form example-seat@example@<host>.
--json JSON output for agents
--no-nudge Suppress the default nudge to the new destination
--note <text> Transition note
--priority <priority> Override priority for the new qitem
--summary <text> OPR.0.4.1.18: short human-readable 1-2 sentence summary for the new qitem — what it is and why this seat, skimmable in the needs-you view (--body stays source of truth). Warned-if-missing.
--tags <tags> Comma-separated tags for the new qitem
--target-repo <name> PL-007: typed repo scope for the new qitem
--tier <tier> Override tier for the new qitem
--to <session> Destination seat receiving the new qitem

rig queue handoff-and-complete

Flag Purpose / help default
--body <text> New qitem body inline (use - to read from stdin; mutually exclusive with --body-file). Omit both to keep the source body.
--body-file <path> Read the new qitem body from a file path (use - for stdin; mutually exclusive with --body). Kills the backtick-shell-corruption class.
--evidence-ref <path> OPR.0.4.4.19 FR-5: durable-artifact pointer for the new qitem. Required by the daemon when the new qitem is human-routed; optional otherwise.
--from <session> (deprecated, ignored) the handing-off seat is derived from the seat env (X-OpenRig-Session); P21 I3 made the handoff route derive it from the transport header
--gate <role> OPR.0.4.3.16: mark the new qitem as gate work; translated to a gate:<role> tag (e.g. guard | spec-review). The idle-gate watchdog reads this predicate. Composes with --tags.
-h, --help display help for command
--host <id> OPR.0.4.6.MH3: route this queue write to a REGISTERED remote host (see rig host ls). EXPLICIT-ONLY — queue verbs never follow the persisted 'rig host select' selection. Equivalent to the host-qualified destination form example-seat@example@<host>.
--json JSON output for agents
--no-nudge Suppress the default nudge to the new destination
--note <text> Transition note
--priority <priority> Override priority for the new qitem
--summary <text> OPR.0.4.1.18: short human-readable 1-2 sentence summary for the new qitem — what it is and why this seat, skimmable in the needs-you view (--body stays source of truth). Warned-if-missing.
--tags <tags> Comma-separated tags for the new qitem
--target-repo <name> PL-007: typed repo scope for the new qitem
--tier <tier> Override tier for the new qitem
--to <session> Destination seat receiving the new qitem

rig queue inbox-absorb

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--receiver <session> Receiver session (must match destination)

rig queue inbox-deny

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--reason <text> Reason for denial
--receiver <session> Receiver session (must match destination)

rig queue inbox-drop

Flag Purpose / help default
--audit <pointer> Audit pointer reference
--body <text> Inbox body inline (use - to read from stdin; mutually exclusive with --body-file).
--body-file <path> Read the inbox body from a file path (use - for stdin; mutually exclusive with --body). Kills the backtick-shell-corruption class.
-h, --help display help for command
--id <inboxId> Idempotent inbox_id
--json JSON output for agents
--sender <session> (deprecated, ignored) sender is derived from the authenticated seat env
--tags <tags> Comma-separated tags
--urgency <urgency> routine | urgent | critical (default: "routine")

rig queue inbox-pending

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig queue list

Flag Purpose / help default
-a, --all Include closed/done history (like 'docker ps -a')
-A, --all-rigs Cross-rig breadth (like 'kubectl get --all-namespaces')
--destination <session> Filter by destination session
--full Show complete per-item fields (body, chain-of-record)
-h, --help display help for command
--json JSON output (compact; use --full --json for complete fields)
--limit <n> Result limit (default: 100)
--mine Scope to items where you are source or destination, including rows you authored but do not own
--owned Scope to obligations assigned to you (destination only)
--source <session> Filter by source session
--state <state> Filter by state (comma-separated for multiple)
--target-repo <name> PL-007: filter qitems by target_repo (exact match)
-o <format> Output format: json

rig queue outbox-list

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--limit <n> Result limit (default: "100")

rig queue outbox-record

Flag Purpose / help default
--audit <pointer> Audit pointer reference
--body <text> Outbox body inline (use - to read from stdin; mutually exclusive with --body-file).
--body-file <path> Read the outbox body from a file path (use - for stdin; mutually exclusive with --body). Kills the backtick-shell-corruption class.
--destination <session> Destination session
-h, --help display help for command
--id <outboxId> Idempotent outbox_id
--json JSON output for agents
--sender <session> (deprecated, ignored) the sender is derived from the seat env (X-OpenRig-Session); P21 I3 made the outbox-record route derive it from the transport header
--tags <tags> Comma-separated tags
--urgency <urgency> routine | urgent | critical (default: "routine")

rig queue overdue

Flag Purpose / help default
-A, --all-rigs Cross-rig breadth (default is current rig only)
--full Include complete per-item fields (body, chain-of-record)
-h, --help display help for command
--json JSON output for agents
--limit <n> Result limit (default: 50)
--rig <name> Scope to a specific rig (default: current rig from OPENRIG_SESSION_NAME)

rig queue resolve

Flag Purpose / help default
--actor <session> (deprecated, ignored) resolver is derived from the seat env (X-OpenRig-Session)
--bearer <token> Operator bearer token for the mission-control write gate (or set OPENRIG_AUTH_BEARER_TOKEN; loopback daemons without a configured bearer need none)
--decision <text> The human's decision text (non-empty; lands in transition_note + the audit row)
-h, --help display help for command
--json JSON output for agents
--no-notify Skip the best-effort owner nudge (the unpark still commits)

rig queue show

Flag Purpose / help default
--full Complete original record; may be large (use --full --json for lossless JSON)
-h, --help display help for command
--json JSON preview with completeness, original byte size and exact full command

rig queue transitions

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig queue unclaim

Flag Purpose / help default
--destination <session> (deprecated, ignored) the releaser is derived from the seat env (X-OpenRig-Session); the unclaim route derives it from the transport header
-h, --help display help for command
--json JSON output for agents
--reason <text> Reason for unclaim (default: "manual")

rig queue undelivered

Flag Purpose / help default
-A, --all-rigs Cross-rig breadth (default is current rig only)
--full Include complete per-item fields (body, chain-of-record)
-h, --help display help for command
--json JSON output for agents
--limit <n> Result limit (default: 50)
--rig <name> Scope to a specific rig (default: current rig from OPENRIG_SESSION_NAME)

rig queue update

Flag Purpose / help default
--actor <session> (deprecated, ignored) the actor is derived from the seat env (X-OpenRig-Session); P21 I3 made the update route derive it from the transport header
--blocked-on <blocker> For state=blocked: the blocker — a qitem id (must exist and be live), a human seat (FR-6 park; requires summary + evidence_ref), or a typed non-qitem gate 'fold:<what>' / 'auth:<what>' / 'external:<what>'
--closure-reason <reason> Required for state=done; also 'superseded' on state=canceled (with --closure-target = the successor) records a supersession, distinct from an abandoned cancel
--closure-target <target> Required for handed_off_to, blocked_on, escalation, and superseded
--evidence-ref <path> OPR.0.4.4.19 FR-6: park-time durable-artifact pointer persisted onto the item (human-seat parks only)
-h, --help display help for command
--json JSON output for agents
--note <text> Transition note for the audit log
--reopen Explicitly acknowledge a deliberate terminal-to-active repair; requires --state and --note
--state <state> New state: pending | in-progress | done | blocked | failed | denied | canceled | handed-off
--summary <text> OPR.0.4.4.19 FR-6: park-time summary persisted onto the item (human-seat parks only)
--wake-after <duration> For state=blocked: atomically arm a timer (for example 90s, 15m, 2h)
--wake-watchdog <jobId> For state=blocked: attach an existing live watchdog id targeting the row owner

rig queue whoami

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--recent-limit <n> How many recent active qitems to include (default: "25")
--session <session> Caller's session name (defaults to OPENRIG_SESSION_NAME)

rig reconcile-session

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--no-launch Never launch/relaunch (the only mode this command has; accepted for explicitness)
--node <logicalId> Disambiguate: target node logical id (requires --rig)
--rig <rigId> Disambiguate: target rig id (requires --node)

rig release

Flag Purpose / help default
--delete Delete the rig record after a clean release
-h, --help display help for command
--json JSON output

rig remove

Flag Purpose / help default
--fallback <live-seat> Reroute active qitems to this running seat before removal
-h, --help display help for command
--json JSON output

rig requirements

Flag Purpose / help default
-h, --help display help for command
--json Output as parseable JSON

rig restore

Flag Purpose / help default
-h, --help display help for command

rig restore-check

Flag Purpose / help default
--full Show complete per-seat detail (today's default)
-h, --help display help for command
--json JSON output (compact summary by default; use --full --json for complete detail)
--no-hooks Skip hook checks
--no-queue Skip queue file checks
--ready Show ready-seat detail in addition to not-ready
--rig <name> Check one rig only

rig restore-packet

Flag Purpose / help default
-h, --help display help for command

rig restore-packet read

Flag Purpose / help default
-h, --help display help for command
--json Emit machine-readable JSON output

rig restore-packet validate

Flag Purpose / help default
-h, --help display help for command
--json Emit machine-readable validation report

rig restore-packet write

Flag Purpose / help default
--authority-boundaries <text> Explicit statement of restored-seat authority
--caveat <text> Caveat to include in the packet (repeatable) (default: [])
--current-work-summary <text> One-paragraph summary of the source seat's current work
--default-target-repo <path> Default target repo absolute path or null
--generator-version <version> Generator version identifier (default: "example-seat@example.1.0")
-h, --help display help for command
--next-owner <name> Who the restored seat hands off to next (default: "self")
--role-pointer <path> Role guidance pointer path or URI for the restored seat
--source-jsonl <path> Source Codex/Claude JSONL transcript file (direct)
--source-rig-override <rig> Operator override for source_rig (R2; required when canonical session id is bare with no <seat>@<rig> shape)
--source-runtime <runtime> Force source runtime (claude-code | codex); auto-detected if omitted
--source-session <session> Source session name (daemon-backed)
--source-session-id-override <id> Operator override for source_session_id (R2; required when JSONL has no session_meta record)
--source-trust-ranking <csv> Comma-separated source-trust ranking; default: rig_whoami,bounded_latest_transcript
--target <dir> Target packet directory (must not exist; created atomically)
--target-rig <rig> Target rig name for the restored seat
--target-runtime <runtime> Target runtime (claude-code | codex)
--target-workspace-root <path> Target workspace root absolute path

rig restore status

Flag Purpose / help default
-h, --help display help for command
--json JSON output
--rig <rigId> Rig ID containing the restore attempt

rig scope

Flag Purpose / help default
-h, --help display help for command
--workspace <path> Override workspace root (otherwise inferred from cwd or $OPENRIG_WORK_ROOT)

rig scope audit

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable JSON output
--mission <name> Mission to audit

rig scope mission

Flag Purpose / help default
-h, --help display help for command

rig scope mission approve

Flag Purpose / help default
--actor <session> (deprecated, ignored) approver is derived from the seat env (X-OpenRig-Session)
-h, --help display help for command
--json Machine-readable output
--locked-artifacts <paths> PLAN-LOCK ONLY (--scope spec): comma-separated slice-relative paths naming the artifact set this lock freezes — replaces the derived default entirely. Each file must exist. Without it, a derivation that would freeze only a missing/scaffold PRD refuses.
--mission <name> (unused at mission tier)
--on-behalf-of <human> Record the delegation: whose decision this stamp records (actor stays the real invoking session)
--re-approve OPR.0.5.0.18 amend/re-stamp: supersede an existing stamp with a new reasoned attestation (prior preserved in the append-only audit log). Requires --reason.
--reason <why> Why the stamp is being amended (required with --re-approve; recorded on the audit row)
--scope <scope> Approval scope: spec | delivery (default delivery)

rig scope mission create

Flag Purpose / help default
--depends-on <dot-id...> Advisory build-order dependencies on sibling mission dot-IDs
-h, --help display help for command
--id <dot-id> Explicit dot-ID. Overrides name-pattern inference.
--intent <text> Authored intent stored in SPEC.md frontmatter (defaults to title)
--json Machine-readable output
--no-mission-notes Deprecated alias for --no-notes
--no-notes Skip NOTES.md scaffolding
--template <kind> Template: placeholder | release (auto when name matches release-X.Y.Z) (default: "")
--title <text> Display title (defaults to titlecased name)

rig scope mission graph

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output

rig scope mission ls

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output

rig scope mission progress

Flag Purpose / help default
--add <text> Append a checkbox row with this text
-h, --help display help for command
--json Machine-readable output
--section <heading> Section heading for --add (default: Rail)
--set <text> Set the status of the row whose trimmed text exactly matches
--status <status> Row status: active | done | blocked (default: "active")

rig scope mission repair

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output

rig scope mission show

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output

rig scope mission stage

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output
--successor <id> Successor scope id — REQUIRED when new-stage is superseded

rig scope mission verified

Flag Purpose / help default
--against <source> What it was verified against — MANDATORY (no bare timestamps)
-h, --help display help for command
--json Machine-readable output

rig scope resolve-notes

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable JSON output

rig scope slice

Flag Purpose / help default
-h, --help display help for command

rig scope slice approve

Flag Purpose / help default
--actor <session> (deprecated, ignored) approver is derived from the seat env (X-OpenRig-Session)
-h, --help display help for command
--json Machine-readable output
--locked-artifacts <paths> PLAN-LOCK ONLY (--scope spec): comma-separated slice-relative paths naming the artifact set this lock freezes — replaces the derived default entirely. Each file must exist. Without it, a derivation that would freeze only a missing/scaffold PRD refuses.
--mission <name> Hint mission when slice-path is just NN-slug
--on-behalf-of <human> Record the delegation: whose decision this stamp records (actor stays the real invoking session)
--re-approve OPR.0.5.0.18 amend/re-stamp: supersede an existing stamp with a new reasoned attestation (prior preserved in the append-only audit log). Requires --reason.
--reason <why> Why the stamp is being amended (required with --re-approve; recorded on the audit row)
--scope <scope> Approval scope: spec | delivery (default delivery)

rig scope slice close

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output
--mission <name> Hint mission when slice-path is just NN-slug
--note <text> Optional closure note
--reason <reason> Closure reason: wontfix | deferred | superseded | stale

rig scope slice create

Flag Purpose / help default
--depends-on <dot-id...> Advisory build-order dependencies on sibling slice dot-IDs
-h, --help display help for command
--intent <text> Authored intent stored in SPEC.md frontmatter (defaults to title)
--json Machine-readable output
--readme-only Write progress_rail: readme-only in README frontmatter instead of scaffolding PROGRESS.md
--template <kind> Template: placeholder | bug-fix | backlog-deprecation | backlog-tech-debt | release-feature | research (default: "placeholder")
--title <text> Display title (defaults to titlecased slug)

rig scope slice ls

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output
--mission <name> Restrict to a single mission
--state <state> Filter: active | closed | shipped | all (default: "active")

rig scope slice move

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output
--mission <name> Hint source mission when slice-path is just NN-slug

rig scope slice progress

Flag Purpose / help default
--add <text> Append a checkbox row with this text
-h, --help display help for command
--json Machine-readable output
--mission <name> Hint mission when slice-path is just NN-slug
--section <heading> Section heading for --add (default: Rail)
--set <text> Set the status of the row whose trimmed text exactly matches
--status <status> Row status: active | done | blocked (default: "active")

rig scope slice repair

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output
--mission <name> Hint mission when slice-path is just NN-slug

rig scope slice ship

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output
--mission <name> Hint mission when slice-path is just NN-slug

rig scope slice show

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output
--mission <name> Hint mission when path is just NN-slug

rig scope slice stage

Flag Purpose / help default
-h, --help display help for command
--json Machine-readable output
--mission <name> Hint mission when slice-path is just NN-slug
--successor <id> Successor scope id — REQUIRED when new-stage is superseded

rig scope slice verified

Flag Purpose / help default
--against <source> What it was verified against — MANDATORY (no bare timestamps)
-h, --help display help for command
--json Machine-readable output
--mission <name> Hint mission when slice-path is just NN-slug

rig seat

Flag Purpose / help default
-h, --help display help for command

rig seat clean

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--operator <address> Operator recorded on the audit event
--reason <text> Audit reason recorded on the session.cleaned event

rig seat clear-attention

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--reason <text> Operator attestation override (skip evidence gate)

rig seat handover

Flag Purpose / help default
--dry-run Plan the handover without changing topology
-h, --help display help for command
--json JSON output for agents
--operator <address> Operator initiating the handover
--reason <reason> Why the handover is happening
--source <source> Source: fresh (default; launches a new agent), discovered:<id> (operator-prepared), fork:<id> (native fork of the source conversation), or rebuild (fresh agent primed from the seat's durable artifacts).

rig seat launch

Flag Purpose / help default
--fresh Explicitly create a blank native occupant; no continuity source is used
-h, --help display help for command
--json JSON output for agents
--operator <address> Operator recorded on the audit event
--reason <text> Audit reason recorded on the seat.fresh_launched event
--stop Stop the current live managed occupant before launching fresh

rig seat set-model

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--model <id> Target model id (e.g. the canonical id an alias pin migrates to)
--operator <address> Operator recorded on the audit event
--reason <text> Audit reason recorded on the node.model_changed event

rig seat set-resume-token

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--reason <text> Operator attestation recorded in the append-only audit event
--token-stdin Read the resume token from STDIN (the only supported input path)

rig seat status

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig seat stop

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--operator <address> Operator recorded on the audit event
--reason <text> Audit reason recorded on the session.stopped event

rig seat switch-client

Flag Purpose / help default
--client <id> Target a specific attached tmux client (required when multiple are attached)
-h, --help display help for command
--json JSON output for agents
--to-window <n> Target window index (default: 0, the canonical seat window)

rig send

Flag Purpose / help default
--context <ref> Deliver a context pack by its path-like ref (e.g. packs/compaction-restore). The resolved content is sent; an oversized ref is flagged as 'walk-sized'.
--dangerously-interact DANGEROUS: deliberately drive an interactive prompt/permission block (implies --raw; requires --reason). The ONLY override of the prompt/permission guard.
--force Back-compat no-op: a mid-task/busy pane already sends-with-advisory by default; --force never bypasses the interactive-prompt/permission guard
--from <session> DEPRECATED + IGNORED (P21 I4). The rendered From:/actor now DERIVES from the transport identity ($OPENRIG_SESSION_NAME, stamped as X-OpenRig-Session) — never a caller-supplied string, which was specimen-5's forgeable surface. Cross-host origin is carried by the relay re-stamping from its authenticated context.
-h, --help display help for command
--host <id> Send on a remote host declared in ~/.openrig/hosts.yaml (ssh hosts shell out; http hosts go CLI-direct to the remote daemon)
--json JSON output for agents
--pod <name> Send to every seat in a pod (fan-out, per-recipient results)
--raw Send exact text/keystrokes without the From/To messaging envelope (still guarded against interactive prompts)
--reason <text> Why the prompt is being driven (required with --dangerously-interact; recorded in the audit log)
--rig <name> Send to every seat in a rig (fan-out, per-recipient results)
--to <sessions> Multi-recipient: comma-list or repeated (--to a,b or --to a --to b) (default: [])
--verify Verify pane only delivery by checking content after send
--wait-for-idle <seconds> Wait until the target is explicitly idle before sending

rig setup

Flag Purpose / help default
--dry-run Show the plan without making changes
--full Install broader operator workstation tools
-h, --help display help for command
--json Machine-readable JSON output
--policy <name> Record a deliberate permission-policy choice into an existing spec (locked|standard|open|yolo|none)
--spec <path> Existing rig spec (file or directory) to record the --policy choice into

rig shrink

Flag Purpose / help default
--fallback <live-seat> Reroute active qitems to this running seat before removing the pod
-h, --help display help for command
--json JSON output

rig skill

Flag Purpose / help default
-h, --help display help for command

rig skill audit

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig skill loadout

Flag Purpose / help default
--apply Apply the reconciled loadout; default is read-only inspection
--cwd <path> Target working directory (default: current directory)
-h, --help display help for command
--json JSON output
--project-root <path> Project root whose project.yaml supplies install.skills (default: cwd)
--runtime <runtime> Target runtime: claude-code or codex
--topology <ids> Comma-separated topology/profile skill identities

rig slack

Flag Purpose / help default
-h, --help display help for command

rig slack disable

Flag Purpose / help default
--actor <actor> Named operator when outside a managed seat
-h, --help display help for command
--reason <reason> Why human delivery is being shut down (recorded in the lifecycle receipt)

rig slack enable

Flag Purpose / help default
--actor <actor> Named operator when outside a managed seat
-h, --help display help for command
--reason <reason> Reason recorded with the change (default: "enable human delivery")

rig slack inbound

Flag Purpose / help default
-h, --help display help for command

rig slack outbound

Flag Purpose / help default
-h, --help display help for command
--json (ignored)

rig slack setup

Flag Purpose / help default
--actor <actor> Named operator when outside a managed seat
--channel <id> Slack channel id the connector app must be a member of
-h, --help display help for command
--inbound-destination <session> where inbound human messages land (default example-seat@example)
--minimum-level-that-interrupts <level> minimum OWNER level that mentions/interrupts: RECORD|NOTICE|ALERT
--minimum-level-that-posts <level> minimum OWNER level posted to Slack: RECORD|NOTICE|ALERT
--reason <reason> Reason recorded with the configuration change (default: "configure human delivery")
--required-scopes <csv> comma-separated bot scopes to require at verify time
--secrets-env-file <path> path to the 0600 env file with SLACK_BOT_TOKEN / SLACK_APP_TOKEN
--source-label <label> label shown in the posted message footer (where the queue lives)

rig slack status

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig slack verify

Flag Purpose / help default
--actor <actor> Named operator when outside a managed seat
-h, --help display help for command
--json JSON output
--reason <reason> Reason recorded with the verification (default: "verify human delivery")

rig snapshot

Flag Purpose / help default
-h, --help display help for command
--intended-seats <ids> Comma-separated intended topology roster stored with the snapshot

rig snapshot list

Flag Purpose / help default
-h, --help display help for command

rig spec

Flag Purpose / help default
-h, --help display help for command

rig spec audit

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig spec preflight

Flag Purpose / help default
-h, --help display help for command
--json JSON output
--rig-root <root> Root directory for pod-aware resolution

rig spec show

Flag Purpose / help default
--as-template Strip instance-specific source state and use a replacement name
-h, --help display help for command
--json JSON output

rig spec validate

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig specs

Flag Purpose / help default
-h, --help display help for command

rig specs add

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig specs ls

Flag Purpose / help default
-h, --help display help for command
--json JSON output
--kind <kind> Filter by kind (rig, agent, or workflow)

rig specs preview

Flag Purpose / help default
-h, --help display help for command
--json JSON output
--kind <kind> Disambiguate by kind (rig, agent, or workflow)

rig specs remove

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig specs rename

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig specs show

Flag Purpose / help default
-h, --help display help for command
--json JSON output
--kind <kind> Disambiguate by kind (rig, agent, or workflow)

rig specs sync

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig start

Flag Purpose / help default
--all Headless: restore all rigs with restore-usable snapshots (zero prompts)
-h, --help display help for command
--json JSON output for agents
--last Headless: restore all rigs that were last running (zero prompts)
--rigs <names...> Headless: restore only the named rigs (zero prompts)

rig startup-proof

Flag Purpose / help default
-h, --help display help for command

rig startup-proof submit

Flag Purpose / help default
--answer <answer> Expected answer from the startup prompt
--challenge-id <id> Startup challenge id from the startup prompt
-h, --help display help for command
--json Print raw JSON response

rig status

Flag Purpose / help default
-h, --help display help for command

rig stream

Flag Purpose / help default
-h, --help display help for command

rig stream archive

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig stream emit

Flag Purpose / help default
--body <text> Stream item body
--format <fmt> Observation body format
-h, --help display help for command
--hint-destination <session> Hint at intended destination seat
--hint-tags <tags> Comma-separated hint tags
--hint-type <type> Hint type (e.g. review, handoff, idea)
--hint-urgency <urgency> Hint urgency (routine, urgent, critical)
--id <streamItemId> Idempotent stream_item_id (skip if not provided)
--interrupt Mark item as interrupting
--json JSON output for agents
--source <session> Source session (e.g. example-seat@example)

rig stream list

Flag Purpose / help default
--after <sortKey> Cursor pagination — return items after this sort key
-h, --help display help for command
--hint-destination <session> Filter by hint destination
--include-archived Include archived items
--json JSON output for agents
--limit <n> Result limit (default: "100")
--since <iso> Include items emitted at or after this ISO timestamp
--source <session> Filter by source session
--tag <tag> Filter by exact hint tag
--until <iso> Include items emitted at or before this ISO timestamp

rig stream show

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig stream watch

Flag Purpose / help default
-h, --help display help for command
--json Emit one StreamItem JSON object per line

rig terminal

Flag Purpose / help default
-h, --help display help for command

rig terminal open

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--provider <name> terminal provider: herdr (default) or cmux (best-effort)

rig terminal status

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--provider <name> restrict to one provider (herdr / cmux)

rig terminal views

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig transcript

Flag Purpose / help default
--grep <pattern> Search for lines matching pattern (regex)
-h, --help display help for command
--host <id> Read from a remote host declared in ~/.openrig/hosts.yaml (http hosts only — CLI-direct to the remote daemon's transcript routes)
--json JSON output for agents
--tail <lines> Show last N lines (default: 50) (default: "50")

rig tui

Flag Purpose / help default
-h, --help display help for command
--shared join the kernel's existing shared terminal; detach with Ctrl-b d

rig tui commands

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig ui

Flag Purpose / help default
-h, --help display help for command

rig ui open

Flag Purpose / help default
-h, --help display help for command

rig unarchive

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig unclaim

Flag Purpose / help default
-h, --help display help for command
--json JSON output

rig up

Flag Purpose / help default
--cwd <path> Override launch working directory for all members for this run only
--existing Treat <source> as an existing rig name; bypass library-spec name resolution
--fresh <seats...> Deliberately fresh-prime the named seats (logical ids) instead of resuming their original sessions (operation B; reported as fresh-primed)
-h, --help display help for command
--host <id> Run on a remote host declared in ~/.openrig/hosts.yaml
--json JSON output for agents
--plan Plan mode — preview without executing
--target <root> Target root directory for package installation (.rigbundle only; does not change agent cwd)
--yes Auto-approve trusted actions

rig usage

Flag Purpose / help default
-h, --help display help for command

rig usage series

Flag Purpose / help default
-h, --help display help for command
--json machine-readable (verbatim route payload)
--lane <lane> context | provider_window
--limit <n> max rows
--seat <session> filter to one seat session
--since <iso> absolute lower bound on captured_at (inclusive)
--until <iso> absolute upper bound on captured_at (exclusive)

rig usage top

Flag Purpose / help default
-h, --help display help for command
--json machine-readable (verbatim route payload)
--top <n> cap the ranking to N seats
--window <duration> look-back window (90m, 1h, 2d, or bare hours) (default: "1h")

rig view

Flag Purpose / help default
-h, --help display help for command

rig view list

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig view register

Flag Purpose / help default
--definition <sql> SQL definition (operator-supplied; not validated for taxonomy)
-h, --help display help for command
--json JSON output for agents
--name <name> Custom view name (must NOT collide with built-in names)
--session <session> Registering operator session

rig view show

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--limit <n> Result row limit (default: "100")
--mission <id> Mission scope for the execution view (defaults to the newest release-* mission)
--rig <rig> Filter by rig name (matches destination_session OR source_session @<rig>)

rig walk

Flag Purpose / help default
--budget <tokens> With --through-profile: situation budget (reported, never truncated)
--consume-poll <duration> Consumption-verification poll interval (e.g. 1500ms or 2s; unit suffix required); default 1500ms
--consume-timeout <duration> Per-piece consumption-verification window (e.g. 20s or 500ms; unit suffix required); default 20s
-h, --help display help for command
--json JSON output
--mission <mission> With --through-profile: the mission-tree grant
--pace <duration> Delay between pieces (e.g. 10s or 500ms; unit suffix required); default 10s
--profile <profile> With --through-profile: named install profile declared by the pack
--rig <rig> With --through-profile: the seat-tree grant (with --seat)
--runtime <runtime> With --through-profile: claude | codex (default claude)
--seat-grant <seat> With --through-profile: the seat whose tree seat: atoms may read (with --rig)
--situation <situation> With --through-profile: fresh | handover | post-compaction
--slice <slice> With --through-profile: the slice-tree grant (with --mission)
--through <items...> A context ref (a pack) OR a list of files to walk the seat through
--through-profile <ref> Walk the seat through a pack's COMPOSED PROFILE (requires --situation; the piece set comes from rig context profile, never hand-authored)
--turn-timeout <duration> Max wait for the seat's turn to CLOSE after a piece is consumed (e.g. 300s; unit suffix required); default 300s

rig watchdog

Flag Purpose / help default
-h, --help display help for command

rig watchdog list

Flag Purpose / help default
-a, --all Include stopped and terminal history
--full Show complete per-job fields
-h, --help display help for command
--json JSON output for agents (compact unless --full)
--limit <n> Result limit (default: 100 unless --full)

rig watchdog register

Flag Purpose / help default
--active-wake-interval-seconds <n> (pool-ready-specific) wake-up cadence when actionable artifacts exist
-h, --help display help for command
--interval-seconds <n> Evaluation interval (positive integer)
--json JSON output for agents
--message <text> (context-usage-threshold) message delivered when the threshold fires
--policy <policy> Policy name (one of: periodic-reminder, artifact-pool-ready, edge-artifact-required, workflow-keepalive, idle-gate-qitem, context-usage-threshold)
--registered-by <session> Registering session (for audit)
--requires-job <jobId> (context-usage-threshold) require this earlier job's receipt for the same occupant
--scan-interval-seconds <n> (pool-ready-specific) artifact pool scan cadence
--spec <path> Path to YAML spec file (optional for context-usage-threshold)
--target-session <session> Canonical <member>@<rig> target
--threshold-bytes <n> (context-usage-threshold) transcript byte threshold
--threshold-mb <n> (context-usage-threshold) transcript threshold in decimal MB
--watched-file <path> (context-usage-threshold) explicit transcript file; recorded context path is the fallback

rig watchdog show

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig watchdog status

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig watchdog stop

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--reason <text> Stop reason (free-form; recorded in terminal_reason)

rig whoami

Flag Purpose / help default
--all-hosts Fan out to all registered HTTP hosts
--full Show the complete whoami payload (contextUsage, commands, runtimeContext, workspace, all sub-fields)
-h, --help display help for command
--host <id> Run on a remote host declared in ~/.openrig/hosts.yaml
--hosts <ids> Fan out to specific hosts (comma-separated)
--json JSON output for agents (compact identity-recovery projection by default)
--node-id <id> Resolve by node ID
--session <name> Resolve by session name
--verbose Alias for --full

rig workflow

Flag Purpose / help default
-h, --help display help for command

rig workflow abort

Flag Purpose / help default
--actor-session <session> Session performing the abort
-h, --help display help for command
--json JSON output for agents
--reason <text> Why the whole instance is being aborted

rig workflow compile

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--operation-key <key> Opaque replay identity to include in the compilation

rig workflow continue

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig workflow guidance

Flag Purpose / help default
--component <id> Choose an effective selected component; never infer a process stage
--full Expand complete relevant prose and selection; preserves authored caveats
-h, --help display help for command
--json Structured guidance and source identities
--packet <id> Select current packet custody when the frontier branches

rig workflow instantiate

Flag Purpose / help default
--created-by <session> Session creating the instance (canonical <member>@<rig>)
--entry-owner <session> Override default entry-step owner
-h, --help display help for command
--json JSON output for agents
--rig <name> Bind the instance to this rig (overrides the spec's target.rig default; roles resolve to seats on this rig)
--root-objective <text> Root objective for the run

rig workflow instantiate-lifecycle

Flag Purpose / help default
--created-by <session> Session creating the instance
--entry-owner <session> Override the compiled entry owner
-h, --help display help for command
--json JSON output for agents
--operation-key <key> Opaque stable identity for exact replay
--rig <name> Bind the instance to this rig
--root-objective <text> Root objective for the run

rig workflow list

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--status <s> Filter by status (active | waiting | completed | failed | aborted)

rig workflow operation

Flag Purpose / help default
-h, --help display help for command
--json Complete original receipt and current instance

rig workflow project

Flag Purpose / help default
--acceptance-candidate <identity> Typed acceptance candidate identity (use with verdict and evidence-ref)
--acceptance-evidence-ref <path> Typed acceptance evidence path (use with candidate and verdict)
--acceptance-verdict <verdict> Typed acceptance verdict (use with candidate and evidence-ref)
--actor-session <session> Session closing the packet (owner-as-author)
--blocked-on <ref> For waiting exits: blocker reference (qitem id, gate name)
--current-packet <qitem-id> qitem being closed (must be on instance frontier)
--evidence-ref <ref> Record attributed progress evidence; a changed reference resets an intentional-wait reminder
--exit <kind> Closure exit kind: handoff | waiting | done | failed
-h, --help display help for command
--instance <id> Workflow instance id
--json JSON output for agents
--next-owner <session> Override default next-step owner
--result-note <text> Closure result note (audit context)
--wait-for-proof <scope> For waiting: watch a slice proof scope through its current attention revision

rig workflow resume

Flag Purpose / help default
--actor-session <session> Session performing the resume (recorded as provenance)
--decision <text> Durable instruction for the step owner (lands in the redrive packet)
-h, --help display help for command
--json JSON output for agents
--occurrence <failed-qitem-id> Failure occurrence to redrive (required when more than one is unresolved)

rig workflow revise

Flag Purpose / help default
--actor-session <session> Agent recording this decision
--apply Apply the inspected compatible proposal (default is read-only)
--expected-digest <sha256> Authored digest returned by inspection
--expected-version <number> Instance version returned by inspection
-h, --help display help for command
--json Complete comparison or receipt
--operation-key <key> Stable revision identity from the inspected apply command
--reason <text> Why the plan changed

rig workflow route

Flag Purpose / help default
--actor-session <session> Session performing the re-route (recorded as provenance)
-h, --help display help for command
--json JSON output for agents
--packet <qitem-id> Frontier packet to route (required when more than one is live)
--reason <text> Why the step is being re-routed (recorded in the audit trail)
--to <session> New owner session (canonical <member>@<rig>)

rig workflow run

Flag Purpose / help default
--created-by <session> Session creating the instance (canonical <member>@<rig>)
--entry-owner <session> Override default entry-step owner
-h, --help display help for command
--json Stream events as JSON lines for agents
--rig <name> Bind the instance to this rig (overrides the spec's target.rig default; roles resolve to seats on this rig)
--root-objective <text> Root objective for the run

rig workflow show

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig workflow specs

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig workflow status

Flag Purpose / help default
-h, --help display help for command
--json Rollup as JSON for agents

rig workflow trace

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig workflow validate

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents

rig workflow watch

Flag Purpose / help default
-h, --help display help for command
--json Stream events as JSON lines for agents

rig workspace

Flag Purpose / help default
-h, --help display help for command

rig workspace doctor

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--strict Exit non-zero on warn-or-fail (default: non-zero only on fail)
--workspace <path> Override the workspace under check (default: daemon-resolved configured root)

rig workspace validate

Flag Purpose / help default
-h, --help display help for command
--json JSON output for agents
--kind <kind> Workspace kind to validate against: user | project | knowledge | lab | delivery
--max-files <n> Hard cap on .md files walked (default: "10000")
--no-recursive Do not descend into subdirectories
--require-frontmatter Report a gap for every .md file without a frontmatter delimiter