# Kirocrew Commands

> Complete CLI reference for Kiro Crew commands. Use for help, commands, setup, how to, what can you do, getting started, onboarding.

- Skill: `kirodotdev/kirocrew-commands` (Agent Skill)
- Install (CLI): `npx skillmds@latest add kirodotdev/kirocrew-commands`
- Raw SKILL.md: https://api.skillmd.com/api/skills/kirodotdev/kirocrew-commands/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: kirodotdev (https://skillmd.com/u/kirodotdev)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/kirodotdev/kirocrew-commands

---

# Kiro Crew CLI Reference

## Setup & System

| Command | Description |
|---------|-------------|
| `kirocrew setup` | Interactive wizard — install agent config (messaging channels connect later) |
| `kirocrew setup --slack` | Also run the guided Slack credential setup (opt-in; ignored with `--agent-only`) |
| `kirocrew setup --agent-only` | Only install kiro-cli agent config, skip the other wizard steps |
| `kirocrew setup --clean` | Fresh install — don't merge from existing config |
| `kirocrew doctor` | Verify Kiro Crew setup (checks all dependencies) |
| `kirocrew doctor --bundle` | Collect logs + crash reports into a redacted diagnostics zip |
| `kirocrew ledger-sweep` | List finished session/work ledgers older than 30 days (dry run, deletes nothing) |
| `kirocrew update` | Update Kiro Crew to the latest version |
| `kirocrew update approve` | Approve a pending in-app update armed from the dashboard |
| `kirocrew update --force` | Discard local commits when a git checkout has diverged from upstream (git installs only) |
| `kirocrew --version` | Print installed version |
| `kirocrew sandbox status` | Report whether this launch is covered by the userns AppArmor profile |
| `kirocrew sandbox install-profile` | Attach the profile to this app (sudo; `--path P` for an explicit executable) |
| `kirocrew sandbox remove-profile` | Unload and remove the profile (sudo) |

`update --force` is destructive on a git install: the hard reset discards local
commits, recoverable only from `git reflog`. The `sandbox` verbs matter only on
hosts with `kernel.apparmor_restrict_unprivileged_userns=1` (Ubuntu 23.10+ and
derivatives) and are no-ops everywhere else; `--path` is refused for
world-writable locations and for shared interpreters such as `/usr/bin/python3`,
which would over-grant.

## Gateway (Server)

| Command | Description |
|---------|-------------|
| `kirocrew gateway` | Start dashboard + Slack gateway |
| `kirocrew gateway --slack-only` | Slack only — skip dashboard web server |
| `kirocrew gateway --no-crons` | Skip cron scheduler |
| `kirocrew gateway --port 9999` | Override dashboard port |
| `kirocrew gateway --port auto` | OS-assigned ephemeral port |
| `kirocrew gateway --no-open` | Don't auto-open dashboard URL in browser |
| `kirocrew gateway --approval reads` | Auto-approve read-only tools |
| `kirocrew gateway --approval yolo` | Auto-approve all tools (requires isolated KIROCREW_HOME) |
| `kirocrew gateway --approval interactive` | Prompt for every tool (default) |
| `kirocrew gateway --seed FIXTURE` | Seed $KIROCREW_HOME from fixture before starting (dev) |
| `kirocrew gateway --seed-replace` | Wipe a non-empty target home and re-seed it (paired with `--seed`) |
| `kirocrew gateway --no-tunnel` | Never publish a tunnel for this process's whole life, whatever `tunnel.enabled` says. Scoped to TUNNELS: it does not change where the dashboard binds |
| `kirocrew gateway --json-ready` | Print one `KIROCREW_READY:{...}` line (port, token, pid, KIROCREW_HOME) once the dashboard is bound |
| `kirocrew gateway --test-mode` | Alias for `--port auto --no-open --json-ready --approval reads` |
| `kirocrew stop` | Stop a running gateway |
| `kirocrew stop --port 9999` | Stop gateway on specific port |
| `kirocrew restart` | Restart gateway (service-aware) |
| `kirocrew status` | Show runtime stats (uptime, sessions, crons, lessons) |

The token in a `--json-ready` line grants dashboard access for up to 20 hours, so
captured stdout from a test harness is a credential: keep it out of logs and
transcripts.

## Service Management

| Command | Description |
|---------|-------------|
| `kirocrew service install` | Install and start as system service (sudo on Linux) |
| `kirocrew service uninstall` | Stop and remove system service |
| `kirocrew service status` | Show service status (systemctl/launchctl) |
| `kirocrew logs` | Show gateway logs (last 100 lines) |
| `kirocrew logs -f` | Follow (tail) live log output |
| `kirocrew logs -n 50` | Show last N lines |

## Pods (Isolated Worktree Test Instances)

Ephemeral, full-stack Kiro Crew gateways — one per feature worktree — that run on
their own port + isolated `KIROCREW_HOME` and never touch the live `:5476`
gateway or shared data. Linux `systemd --user` only. `<wt>` is a worktree name
(resolved by directory basename or `feat/<name>` branch convention).

| Command | Description |
|---------|-------------|
| `kirocrew pod install` | Lay down the systemd --user template unit (once per machine) |
| `kirocrew pod provision <wt>` | Build the worktree's venv + SPA dist (the on-ramp) |
| `kirocrew pod up <wt>` | Bring up an isolated pod (auto-builds venv; fails if dist missing) |
| `kirocrew pod up <wt> --provision` | Provision (venv + dist build) then bring up |
| `kirocrew pod up <wt> --json` | Bring up and print `{base_url, token, port}` as JSON |
| `kirocrew pod ls` | List running pods |
| `kirocrew pod status <wt>` | Up/down + health for one pod |
| `kirocrew pod token <wt>` | (Re)mint a dashboard token for a running pod |
| `kirocrew pod url <wt>` | Print the pod's base URL |
| `kirocrew pod logs <wt> -n N` | Tail the pod's journal |
| `kirocrew pod down <wt>` | Evict the pod and delete its isolated HOME |
| `kirocrew pod prune` | Bulk-reclaim orphaned pod HOMEs (`--older-than 3d` by default, `--all`, `--dry-run`, `--json`) |
| `kirocrew pod scenarios` | List the seed scenarios `pod up --seed <scenario>` accepts (`--json`) |
| `kirocrew pod exec <wt> -- <args>` | Run a kirocrew command against a pod, using the pod's own binary and data |
| `kirocrew pod api <wt> <METHOD> <path>` | Call a running pod's HTTP API with its own token; prints `{name, method, path, status, ok, body}` |
| `kirocrew pod api <wt> POST config --data '{…}' --allow-write` | GET and HEAD are permitted by default; every other method needs `--allow-write` |

**Platform:** Linux only. On macOS/Windows every systemd-touching verb refuses
with a one-line message pointing at `./dev-backend.sh` — it does not crash, and
`pod install` writes no unit file. `pod url` works anywhere (pure computation).

Port derivation: `base + (cksum(name) % 199) + 1` (base `7810` → `7811..8009`).
Override with `PORT=` in `~/.kiro/crew/pods/<name>.env`.

`kirocrew pod --help` lists every verb and its flags.

## Dashboard Access

| Command | Description |
|---------|-------------|
| `kirocrew token` | Print a dashboard URL with auth token (TTL: 20h) |
| `kirocrew token --ttl 1h` | Token with custom TTL (e.g. 1h, 30m) |
| `kirocrew logout` | Revoke all active dashboard sessions |
| `kirocrew manifest` | Generate Slack app manifest with your alias |
| `kirocrew manifest --url` | Print one-click Slack app creation URL |

## Chat

| Command | Description |
|---------|-------------|
| `kirocrew chat` | Interactive chat (REPL mode) |
| `kirocrew chat -m "message"` | Single message (non-interactive) |
| `kirocrew chat --model claude-opus` | Use specific model |

## Browsing (`browser` MCP tool, then `playwright-cli`)

Browsing is not a `kirocrew` subcommand. Start with the **`browser` MCP tool**
in the native Browser panel:
`op=navigate|snapshot|click|type|press_key|hover|select_option|screenshot|wait_for|back|console`.
It refuses loopback, private, and link-local targets. Use `playwright-cli` when
no native panel serves the session, or for attached logins, saved storage state,
and the full operate verbs. It must be on `PATH`; install via **Settings → Browser**
(also holds the attach token) or `npm install -g @playwright/cli@latest`
(Node.js 20 or newer).

| Command | Description |
|---------|-------------|
| `playwright-cli open <url>` | Open a page (prints URL, title, and a snapshot path) |
| `playwright-cli snapshot` | Write the accessibility tree to a YAML file, print its path |
| `playwright-cli click <ref>` / `fill <ref> <text>` | Act on an element from a snapshot |
| `playwright-cli screenshot [ref]` | Write a PNG, print its path. `[ref]` is an ELEMENT, not a path; do not pass `--filename` (it resolves against the CWD and is not auto-approved) |
| `playwright-cli state-save` / `state-load <file>` | Save or restore a logged-in session. Bare `state-save` writes into the service's own directory; both a name and `state-load` prompt for approval, because each names a local path |
| `playwright-cli attach --extension` | Drive the user's own running Chrome, with their logins |
| `playwright-cli show --port <n> --host 127.0.0.1` | Serve the CLI's dashboard for the Browser panel |

**Browsing workflow:** load the `web-browse`, `web-verify`, or `browser-auth`
skill for the task. On the CLI path:
1. Check `command -v playwright-cli`; if absent, use `web_fetch` and give the
   install command above.
2. Run `playwright-cli open <url>`; read its snapshot YAML only when the tree is
   needed, such as before clicking. Refs (`[ref=e5]`) expire on page changes:
   re-snapshot before the next action.
3. On a login redirect, `state-load` a saved session or let the user sign in in
   the Browser panel, then `state-save`.

**No npm access (internal registry, air-gapped host):** detection is **PATH-based**
-- `playwright-cli` on `PATH` is all that matters, so ANY install route works and the
Settings button is a convenience, not the only one. In order of likelihood:

1. Most internal registries proxy npmjs, so the plain install already works.
2. Force the public registry for this one package:
   `npm install -g @playwright/cli --registry=https://registry.npmjs.org`.
3. **An internal registry that requires a login the user does not have** (the
   common Amazon-internal / corporate case). Install into a user-owned prefix
   against the public registry, ignoring the corporate `.npmrc` for this one
   command, then put the binary on `PATH`:

   ```bash
   NPM_CONFIG_USERCONFIG=/dev/null \
     npm install --prefix ~/.local/share/playwright-cli \
     --registry=https://registry.npmjs.org @playwright/cli@0.1.18
   mkdir -p ~/.local/bin
   ln -sf ~/.local/share/playwright-cli/node_modules/.bin/playwright-cli \
     ~/.local/bin/playwright-cli
   ```

   Two caveats worth stating to the user rather than burying: `~/.local/bin` has
   to be **on `PATH`** or Kiro Crew still reports "not installed" (detection is
   `PATH` + the Node bin dirs, nothing else); and `NPM_CONFIG_USERCONFIG=/dev/null`
   deliberately ignores their employer's registry configuration, which is their
   call to make, not ours to assume.
4. Air-gapped: `npm pack @playwright/cli` on a connected machine, copy the
   `.tgz` over, then `npm install -g ./playwright-cli-<version>.tgz`. Note the
   tarball alone is not runnable -- it needs its `playwright` /
   `playwright-core` dependencies resolved too.

What does **not** substitute for it: `pip install playwright` and
`dotnet tool install Microsoft.Playwright.CLI` install a DIFFERENT tool -- the
`playwright` browser-installer/codegen CLI, not `@playwright/cli` (binary
`playwright-cli`, its own 0.x line, which depends on `playwright@1.63.0-alpha`).
Switching to yarn, pnpm or bun hits the same registry, so it only helps when the
`npm` client itself is missing. And there is **no standalone binary**: the
upstream GitHub release carries no build assets and `playwright-cli.js` starts
with `#!/usr/bin/env node`, so Node.js 18+ is required no matter how it is
fetched.

**Approval:** page-scoped verbs run without prompting the user, because installing
the CLI is itself the consent. Verbs that reach the local machine still prompt on
purpose -- `eval`, `run-code`, `upload`, `state-load`, a named `state-save`, and the
installers. Let the user approve those rather than rewriting the command to dodge
the prompt.

The full verb list is in the skill `playwright-cli install --skills agents --global`
writes.

## Autonomous Task Runner

| Command | Description |
|---------|-------------|
| `kirocrew run TASK.md` | Run a task spec file (auto-resumes from checkpoint) |
| `kirocrew run TASK.md --fresh` | Start from scratch, ignore checkpoint |
| `kirocrew run TASK.md --no-test` | Skip build/test verification after each step |
| `kirocrew run TASK.md --timeout 3600` | Set global timeout in seconds |
| `kirocrew run TASK.md --name "My Task"` | Override human-readable task name |

## Subagents

| Command | Description |
|---------|-------------|
| `kirocrew spawn run "task"` | Spawn a background subagent (wait for result) |
| `kirocrew spawn run --async "task"` | Fire-and-forget subagent |
| `kirocrew spawn list` | List active subagents |

## Cron Jobs

| Command | Description |
|---------|-------------|
| `kirocrew cron list` | List all cron jobs |
| `kirocrew cron add NAME MESSAGE --every 3600` | Add job with interval (seconds) |
| `kirocrew cron add NAME MESSAGE --cron "0 9 * * MON-FRI"` | Add job with cron expression |
| `kirocrew cron add NAME MESSAGE --agent myagent` | Add job for specific agent |
| `kirocrew cron add NAME MESSAGE --approval-mode auto` | Add job with auto tool approval |
| `kirocrew cron add NAME MESSAGE --channel C123456` | Post results to Slack channel |
| `kirocrew cron update JOB_ID --message "new msg"` | Update job message |
| `kirocrew cron update JOB_ID --agent myagent` | Update job agent |
| `kirocrew cron update JOB_ID --approval-mode auto` | Set auto-approval |
| `kirocrew cron update JOB_ID --approval-mode default` | Reset approval to default |
| `kirocrew cron remove JOB_ID` | Remove a cron job |
| `kirocrew cron pause JOB_ID` | Pause a cron job |
| `kirocrew cron resume JOB_ID` | Resume a paused job |
| `kirocrew cron adopt JOB_ID --session-of SESSION` | Give the job an owning chat session, so that session manages it and receives its results |
| `kirocrew cron adopt JOB_ID --release` | Clear the owning session, returning the job to CLI/dashboard-only management |
| `kirocrew cron trigger JOB_ID` | Trigger a job immediately |
| `kirocrew cron preview SCRIPT` | Run a script cron locally with real MCP tools; notifications are printed, not delivered |
| `kirocrew cron preview SCRIPT -m "msg" -e K=V` | Preview with an input message / extra env vars |

## Learning & Memory

| Command | Description |
|---------|-------------|
| `kirocrew learn list` | List all saved lessons |
| `kirocrew learn add "rule text"` | Save a lesson (category: knowledge) |
| `kirocrew learn add "rule text" --category tool` | Save with category (tool/preference/knowledge) |
| `kirocrew learn add "rule text" --negative "avoid X"` | Save with negative example |
| `kirocrew learn remove "query"` | Remove lessons matching substring |
| `kirocrew memory list` | Show semantic memory entries |
| `kirocrew memory search "query"` | Search episodic memories |
| `kirocrew memory stats` | Show memory statistics |
| `kirocrew memory audit` | Scan memory for suspicious content |
| `kirocrew memory export` | Export all memory to JSON (stdout) |
| `kirocrew memory export -o file.json` | Export to file |
| `kirocrew memory import file.json` | Import memory from JSON |
| `kirocrew memory migrate` | Migrate legacy markdown memory to vector store |
| `kirocrew memory show [preferences\|projects\|history]` | Show the markdown memory layer (default: all three; `--format md\|json`, `--since YYYY-MM-DD` for history) |
| `kirocrew knowledge dedup` | Preview cross-source duplicate knowledge documents (dry-run) |
| `kirocrew knowledge dedup --apply` | Actually collapse the duplicates |
| `kirocrew consolidate` | List sessions with unconsolidated messages |
| `kirocrew consolidate SESSION_KEY` | Force consolidate a session (triggers auto-skill extraction) |
| `kirocrew consolidate --all` | Consolidate all pending sessions |

## Artifacts

LLM-generated UI components (widgets, HTML, markdown, SVG, JSON, text).

| Command | Description |
|---------|-------------|
| `kirocrew artifact list` | List all artifacts |
| `kirocrew artifact list --tag ops --kind widget` | Filter by tag and kind |
| `kirocrew artifact list -q "CR"` | Substring filter on name |
| `kirocrew artifact show SLUG` | Print artifact content |
| `kirocrew artifact show SLUG --version 2` | Show specific version |
| `kirocrew artifact show SLUG --meta` | Show metadata as JSON |
| `kirocrew artifact save --name "My Widget" --content-file widget.html` | Save new artifact |
| `kirocrew artifact save --name "X" --content "..." --tags ops,cr` | Save with inline content |
| `kirocrew artifact update SLUG --content-file widget.html` | Update artifact content |
| `kirocrew artifact update SLUG --name "New Name" --tags ops` | Rename/retag |
| `kirocrew artifact versions SLUG` | List version numbers |
| `kirocrew artifact delete SLUG` | Delete artifact and all versions |

## Agents & Workspaces

| Command | Description |
|---------|-------------|
| `kirocrew agent list` | List Kiro Crew agents |
| `kirocrew agent create --name NAME` | Create a new agent |
| `kirocrew agent create --name NAME --kiro-agent kirocrew --workspace default` | Full options |
| `kirocrew agent update NAME --kiro-agent new-agent` | Update agent settings |
| `kirocrew agent delete NAME` | Delete an agent |
| `kirocrew agent reset-model [--agent kirocrew]` | Clear a pinned model so the agent tracks the shipped default |
| `kirocrew workspace list` | List workspaces |
| `kirocrew workspace create --name NAME --dir DIRNAME` | Create workspace (`--dir` is a **name under the data home**, not an absolute path) |
| `kirocrew workspace create --name NAME --copy-from existing` | Copy from existing |
| `kirocrew workspace update NAME --dir DIRNAME` | Update workspace dir (same containment rule) |
| `kirocrew workspace delete NAME` | Delete workspace |

**On the CLI**, `--dir` must resolve to a **strict descendant** of
`$KIROCREW_HOME` (default `~/.kiro/crew`): anything landing outside — `/tmp/x`,
`../x`, `~/x` — is refused with a SEL `denied` audit event, and so is the data
home **root itself** (in any spelling: absolute, `~/.kiro/crew`, `.`, or empty),
since a workspace there would put agent-writable memory on top of `config.json`
and `.env`. The test is containment, not "is it absolute": an absolute path
landing *under* the home is accepted, since it resolves where the relative form
would. Pass `workspace-myproject`, not `/path/to/dir`.

Note the surface difference: the **dashboard** `POST /api/workspaces` DOES accept
an absolute `dir` (screened by `is_sensitive_path`, so `~/.ssh` / `~/.aws` /
keystone paths are still refused). The CLI is the stricter of the two.

## Apps

| Command | Description |
|---------|-------------|
| `kirocrew app list` | List installed apps |
| `kirocrew app install /path/to/app-dir` | Install app from local directory (needs app.json) |
| `kirocrew app enable NAME` | Enable an installed app |
| `kirocrew app disable NAME` | Disable an installed app |
| `kirocrew app uninstall NAME` | Uninstall an app and preserve its data directory |
| `kirocrew app uninstall NAME --purge-data` | Uninstall and explicitly delete the app data directory |
| `kirocrew app info NAME` | Show app details |
| `kirocrew app init NAME` | Scaffold a new app (kebab-case name) |
| `kirocrew app init NAME --backend --ui --cron` | Scaffold with backend, UI, and sample cron |
| `kirocrew app dev NAME` | Toggle an app into dev mode (no-store UI serving + live reload on file change) |
| `kirocrew app dev NAME --off` | Leave dev mode |
| `kirocrew app mcp NAME` | Run an app's MCP server on stdio (spawned by kiro-cli, not for humans) |

## Configuration

| Command | Description |
|---------|-------------|
| `kirocrew config get` | Show all config |
| `kirocrew config get agent.provider` | Get specific value (dot-separated key) |
| `kirocrew config set dashboard.url http://localhost:5476` | Set a config value (port is the KIROCREW_PORT env var, not a config key) |
| `kirocrew config set --file config.json` | Load full config from JSON file |
| `kirocrew config edit` | Open config in $EDITOR |
| `kirocrew config defaults [KEYS…]` | Review stored values that still hold a superseded default |
| `kirocrew config defaults [KEYS…] --adopt` | Remove those stored keys so the current defaults apply |
| `kirocrew config defaults [KEYS…] --keep` | Record the stored values as intentional and stop reporting them |

## Profiling (debug-only)

Off unless `KIROCREW_DEBUG=1` is set; the CLI is the only entry point. Emits folded
stacks (open in speedscope / flamegraph.pl). See `docs/architecture/design-notes/profiling.md`.

| Command | Description |
|---------|-------------|
| `KIROCREW_DEBUG=1 kirocrew perf sample --call mod:fn` | Profile that callable in-process (no extra dependency) |
| `KIROCREW_DEBUG=1 kirocrew perf sample` | Attach to the running gateway (needs `pip install "kirocrew[perf]"`) |
| `KIROCREW_DEBUG=1 kirocrew perf sample --pid 1234 --seconds 30` | Attach to a specific PID for N seconds (1-300) |
| `... --interval 0.002` | Seconds between samples (0.001-1.0, default 0.005) |
| `... --output /tmp/p.folded` | Where to write the profile (default `./kirocrew-profile.folded`) |
| `KIROCREW_DEBUG=1 kirocrew desktop metrics` | Per-process CPU/memory of the **Electron** app (`--json`, `--top N`, `--path`) |

On macOS the attach path additionally needs elevated privileges (the OS denies
`task_for_pid`), so it may require sudo; `--call` needs neither py-spy nor sudo.

`desktop metrics` reads a recording rather than querying the app: `getAppMetrics()`
is Electron-main-only, so the app samples itself into an artifact when **started**
with `KIROCREW_DEBUG` set. Setting the variable only for the CLI does not make an
already-running app record -- restart it.

## Security & Eval

| Command | Description |
|---------|-------------|
| `kirocrew secrets import` | Dry-run the migration of plaintext `.env` credentials into the encrypted vault |
| `kirocrew secrets import --apply` | Store the secrets and rewrite `.env` to `secret://` refs |
| `kirocrew telemetry status` | Show exactly what the anonymous beacon sends, and whether it will |
| `kirocrew telemetry disable` | Turn the anonymous beacon off permanently |
| `kirocrew telemetry enable` | Turn the anonymous beacon back on |
| `kirocrew security audit` | Scan conversation history for suspicious tool usage |
| `kirocrew security deny-list` | Show active deny patterns |
| `kirocrew security events` | Show recent security event log entries (last 20) |
| `kirocrew security events -n 50` | Show N entries |
| `kirocrew security verify` | Verify security event log HMAC integrity |
| `kirocrew eval` | Run smoke test evaluation (~30s) |
| `kirocrew eval memory_recall_basic` | Run specific scenario by name |
| `kirocrew eval --all` | Run all scenarios (slow) |
| `kirocrew eval --judge` | Enable LLM judge scoring |
| `kirocrew bench list` | Show the available corpora and what is cached |
| `kirocrew bench fetch CORPUS` | Download a corpus into the local cache and verify its checksum |
| `kirocrew bench retrieval` | Measure retrieval recall/nDCG against a corpus (deterministic) |
| `kirocrew bench kb-retrieval` | Measure Knowledge Library recall/MRR/nDCG against a golden set (deterministic) |
| `kirocrew bench compare A B` | Diff two saved JSON reports; refuses to attribute a delta when the runs disagree on corpus |

## Governance Policy

Inspects the two-level security model (`effective = POLICY ∩ PROFILE`,
tightest-wins). Six verbs: five are read-only, and `fetch` writes — it applies the
central policy when the download is usable. The enterprise ceiling is never
hand-edited through the CLI (its files are keystone-fenced so the agent cannot read
or write them).

| Command | Description |
|---------|-------------|
| `kirocrew policy show` | Show the effective enterprise security policy |
| `kirocrew policy show --ids` | List each denied-command category's rule ids (default: counts only) |
| `kirocrew policy validate` | Load-check the policy + all profiles |
| `kirocrew policy explain SCOPE ITEM` | Explain one tool/scope decision for a surface |
| `kirocrew policy explain SCOPE ITEM --session-key K --agent A --app APP` | Scope the explanation to a surface |
| `kirocrew policy profile NAME` | Show a profile by name |
| `kirocrew policy source` | Show whether this host fetches its policy from a central source |
| `kirocrew policy fetch` | Fetch the central policy now and apply it if usable (`--force` re-downloads, ignoring cached validators) |

## Cloud (Bring-Your-Own AWS)

Runs Kiro Crew on an EC2 instance in **your own** AWS account; credentials are
resolved by the `aws` CLI and never stored by Kiro Crew. All verbs accept
`--profile` / `--region`; the single-instance verbs also accept `--tag`
(defaults to the last launched instance).

| Command | Description |
|---------|-------------|
| `kirocrew cloud doctor` | Check cloud prerequisites + AWS reachability |
| `kirocrew cloud launch` | Provision + configure an instance (interactive) |
| `kirocrew cloud launch --size TIER -y` | Non-interactive launch at a size tier |
| `kirocrew cloud launch --new` | Create a separate new instance instead of resuming the saved one |
| `kirocrew cloud launch --keep-on-failure` | On bootstrap failure keep the instance for inspection |
| `kirocrew cloud list` | List your Kiro Crew cloud instances |
| `kirocrew cloud status` | Show one instance's state |
| `kirocrew cloud connect` | Open the dashboard over an SSM tunnel |
| `kirocrew cloud tunnel` | Open the dashboard SSM tunnel (standalone alias of connect) |
| `kirocrew cloud connect --local-port N --no-browser` | Forward to a specific local port, no browser |
| `kirocrew cloud login` | Sign kiro-cli in on the instance (fixes "not logged in" chat errors) |
| `kirocrew cloud logout` | Sign kiro-cli out on the instance, to switch Kiro account |
| `kirocrew cloud stop` | Stop the instance (pause billing) |
| `kirocrew cloud start` | Start a stopped instance |
| `kirocrew cloud destroy` | Remove the instance and ALL its AWS resources |
| `kirocrew cloud destroy --dry-run` | Show the delete command without running it |
| `kirocrew cloud iam-policy` | Print the least-privilege IAM policy to apply |
| `kirocrew cloud iam-boundary` | Pre-create the immutable permissions boundary (admin, one-time) |

## Tailnet (Tailscale)

Publishes this dashboard on your tailnet and trusts its origin, so a device on the
tailnet reaches it without a public tunnel.

| Command | Description |
|---------|-------------|
| `kirocrew tailnet status` | Show whether the dashboard is published and trusted on your tailnet |
| `kirocrew tailnet up` | Publish the dashboard on your tailnet and trust its origin |
| `kirocrew tailnet down` | Stop publishing the dashboard on your tailnet |
| `... --port N` | Name the dashboard port; `up` needs it whenever discovery has no verified port |

`up` publishes only a port it has evidence for: an explicit `--port`, `KIROCREW_PORT`,
or the running gateway's run marker. With none of those — the gateway is down, the
marker is unreadable, or several gateways are up, where the marker deliberately
refuses — `up` refuses rather than falling back to the configured `dashboard.url`
port, because nothing is verified to answer there and `tailscale serve` would expose
whatever does. Start the gateway and re-run, or name the port yourself. `status` and
`down` do accept the configured port: one only reports, and the other checks mount
ownership before removing anything.

## Computer Use (Desktop Automation)

Default-OFF behind a keystone enable (`~/.kiro/crew/computer_use.json`, **not**
`config.json`). macOS and Windows carry the full tool set; Linux answers a typed
unsupported refusal. On Windows there is no per-process input, so a keystroke takes
the user's keyboard focus and a coordinate click moves their real cursor — say so
rather than reporting a silent success. These are human debug/diagnostic twins of the
`computer_*` MCP tools — the agent uses the MCP tools, not these.

| Command | Description |
|---------|-------------|
| `kirocrew computer doctor` | Report platform support, keystone enable state, and the advisory Accessibility / Screen Recording probe |
| `kirocrew computer doctor --json` | Same as JSON |
| `kirocrew computer apps` | List on-screen applications the accessibility layer can address |
| `kirocrew computer call TOOL k=v …` | Run ONE computer-use tool through the same gated chokepoint the agent uses |
| `kirocrew computer call --calls '[…]'` | Run a JSON array of calls in a SINGLE process, so `element_index` values stay resolvable |

## Snapshot & Restore

| Command | Description |
|---------|-------------|
| `kirocrew snapshot` | Create a portable backup of Kiro Crew state |
| `kirocrew snapshot /path/to/dir` | Snapshot to specific output directory |
| `kirocrew snapshot --keep 7` | Keep N most recent snapshots (default: 7) |
| `kirocrew snapshot --list` | List existing snapshots |
| `kirocrew restore` | Restore from most recent snapshot |
| `kirocrew restore /path/to/snap.tar.gz` | Restore from specific snapshot |
| `kirocrew restore --mode replace` | Replace mode (default) |
| `kirocrew restore --mode merge` | Merge mode |
| `kirocrew restore --dry-run` | Preview without applying |
| `kirocrew restore --components memory,crons` | Restore specific components only |
| `kirocrew restore --list-components` | List restorable components |
| `kirocrew restore --force` | Restore even if gateway is running |

## Slack Commands

### All Allowed Users
| Command | Description |
|---------|-------------|
| `!dashboard` | Get a presigned dashboard link (DM'd to you). Link expires in 5 min; session lasts 1h |
| `!dashboard 2h` | Dashboard link with custom duration (accepts `<N>h` or `<N>m`, max 6h) |
| `/kirocrew dashboard` | Same via slash command |
| `/kirocrew help` | List available slash sub-commands |
| `!stop` | Force-halt the current agent turn (bypasses semaphore, cancels active task) |
| `status` | Show runtime stats |
| `ping` | Auto-reply `pong` |
| `cron list` | List cron jobs |
| `run <path>` | Run an autonomous task from a spec file |

### Owner-Only Slash Commands
| Command | Description |
|---------|-------------|
| `/kirocrew yolo` | Toggle YOLO mode (auto-approve all tool calls) |
| `/kirocrew agent` | Show agent selector dropdown |
| `/kirocrew agent <name>` | Switch to named agent |
| `/kirocrew voice` | Open TTS voice settings modal |
| `/kirocrew config` | Open config modal |
| `/kirocrew users` | Open allowed users management modal |
| `/kirocrew channels` | Open tracked channels modal |
| `/kirocrew sessions` | List recent sessions with resume/end buttons |

## Environment Variables

| Variable | Purpose | Default |
|----------|---------|---------|
| `KIROCREW_HOME` | Override config/data directory | `~/.kiro/crew` |
| `KIROCREW_PORT` | Override dashboard port | `5476` |
| `KIROCREW_PROJECT_DIR` | Override agent config/skills directory | Auto-detected |
| `KIROCREW_POD_REPO` | Repo to resolve worktree names from | invoking cwd |
| `KIROCREW_POD_ROOT` | Isolated pod HOMEs (nuked on stop) | `~/.kirocrew-pods` |
| `KIROCREW_POD_BASE_PORT` | Port derivation base | `7810` |
| `KIROCREW_POD_LIVE_PORT` | Port a pod must never bind | `5476` |