Gitmoot Agent Skill

This root SKILL.md is kept as a raw compatibility entrypoint for agents and
gitmoot.io/SKILL.md. The canonical Agent Skills package lives at
skills/gitmoot/, with deeper reference files under skills/gitmoot/references/.

Gitmoot is a local-first coordinator for AI agents working across repositories,
goals, reviews, PR comments, and runtime workflows. Use this skill when the
user wants PR-comment agent workflows, repo-scoped agent subscriptions,
background daemon checks, Codex, Claude Code, or Kimi Code agent startup,
structured implementation plans, standard goal files, agent template workflows,
template capture, custom prompt agents, job status, or branch lock inspection.

For current-chat prompt import, "use here" means run
gitmoot agent prompt <agent> and apply the returned prompt content in this
current chat. This is prompt import, not true system-prompt injection. Do not
route a "here" request through a background gitmoot agent ask unless the user
explicitly asks for background execution, PR-comment routing, or job tracking.

For fast planning, "use the Gitmoot planner here" is the natural-language
shortcut for gitmoot agent prompt planner. If the planner template is not
cached, read and apply the packaged skills/gitmoot/agent-templates/planner.md
instructions directly.

For template capture, phrases like "capture this session as a Gitmoot agent
template", "turn this workflow into a Gitmoot template", or "draft a reusable
agent template from this chat" mean read
skills/gitmoot/references/TEMPLATE_CAPTURE.md and distill the visible
current-chat context into a draft template. Gitmoot cannot read hidden model
memory or runtime internals. Do not install, overwrite, or update a permanent
template unless the user explicitly approves that step.
Use gitmoot agent template draft <id> for a blank scaffold,
gitmoot agent template validate <file> for a structural check,
gitmoot agent template add <id> --file <file> to install a snapshot, and
gitmoot agent prompt <id> to reuse the installed template in the current chat.

For background work, keep Gitmoot's resource model explicit: repo checkout
locks protect local checkouts, runtime session locks serialize delivery for the
same Codex, Claude, or Kimi session, and branch locks protect implementation
ownership.
The daemon default is --workers 1; raise it only for independent runtime
sessions or managed agent types with max_background greater than one.

For runtime selection, gitmoot agent start <name> --runtime <runtime> accepts
codex, claude, kimi, or kimi-cli (the opt-in legacy Kimi CLI adapter).
Kimi Code is a first-class runtime adapter alongside Codex and Claude Code. To
use it, run kimi login, then restart the Gitmoot daemon so it inherits the
session. agent subscribe additionally accepts --runtime shell, whose
session is a command (for deterministic tests).

For Gitmoot health or status questions, run the relevant read-only Gitmoot CLI
checks and answer directly from the results. Mention gitmoot dashboard only
after that answer, as a live monitoring follow-up. Do not start daemons, create
agents, change subscriptions, update templates, or release locks unless the user
asks for that action.

Before Acting

1. Check whether gitmoot is installed with gitmoot version.
2. Confirm GitHub CLI access with gh auth status before using PR workflows.
3. Detect or ask for the target repo before starting daemons, subscribing agents,
   or routing jobs.
4. Do not start daemons, create agents, update agent templates, change
   subscriptions, or release locks unless the user asks or the current task
   clearly requires it.
5. Prefer read-only status commands and answer directly before mutating Gitmoot
   state or pointing the user to live monitoring.

Install And Update

If gitmoot is not installed, install the latest beta:

curl -fsSL https://gitmoot.io/install.sh | sh

Verify the install and GitHub access before using Gitmoot:

gitmoot version
gitmoot update --check
gh auth status

Install the separate Python optimizer before optimizer-backed SkillOpt train
continues:

python3 -m pip install --user pipx
python3 -m pipx ensurepath
pipx install https://github.com/jerryfane/gitmoot-skillopt/releases/download/v0.3.1/gitmoot_skillopt-0.3.1-py3-none-any.whl
gitmoot-skillopt --version
gitmoot-skillopt optimize --help

If pipx is unavailable, install gitmoot-skillopt in a venv and pass
--skillopt-bin /path/to/venv/bin/gitmoot-skillopt to gitmoot skillopt train continue.

Install runtime plugin guidance when the user wants Codex or Claude Code to
discover Gitmoot from its plugin system:

gitmoot plugin install codex
gitmoot plugin install claude
gitmoot plugin doctor

Common Local Commands

gitmoot status --repo owner/repo
gitmoot events --repo owner/repo
gitmoot daemon start --poll 30s
gitmoot daemon start --session <root-job-id>
gitmoot daemon start
gitmoot daemon status
gitmoot daemon restart
gitmoot daemon stop [--forget-runtime-auth]
gitmoot plugin doctor
gitmoot agent list
gitmoot agent doctor <agent>
gitmoot agent prompt <agent-or-template>
gitmoot agent run <agent> --repo owner/repo "question, review, or implementation request"
gitmoot agent review <agent> --repo owner/repo --pr <number> "review request"
gitmoot agent implement <agent> --repo owner/repo --task <task-id> "implementation request"
gitmoot agent ask <agent> --repo owner/repo "question or instructions"
gitmoot job list --repo owner/repo
gitmoot job show <job-id>
gitmoot job events <job-id>
gitmoot report bug --job <job-id> --preview
gitmoot report bug --job <job-id> --create --yes
gitmoot lock list --repo owner/repo
gitmoot lock show owner/repo <branch>
gitmoot skillopt export --run <run-id> [--output training.json]
gitmoot skillopt import --file candidate.json [--artifact-dir artifacts]
gitmoot skillopt candidate list [--template id]
gitmoot skillopt candidate show <version-id>
gitmoot skillopt candidate promote <version-id>
gitmoot skillopt candidate reject <version-id> [--reason text]
gitmoot skillopt feedback markdown export --run <run-id> --output .gitmoot/evals/<run-id>
gitmoot skillopt feedback markdown import --packet .gitmoot/evals/<run-id>
gitmoot skillopt feedback github publish --run <run-id> [--repo owner/repo] [--pr <number>]
gitmoot skillopt feedback github sync --run <run-id> [--repo owner/repo] (--issue <number>|--pr <number>)

Use gitmoot daemon start for the background daemon. Use gitmoot daemon run
only when the user explicitly wants a foreground process. --repo owner/repo
SCOPES the daemon to a single repo: it polls only that repo's PRs and claims
only that repo's queued jobs. Omit --repo to supervise every enabled
registered repo from one daemon (#581). Both daemon run and daemon start accept --session <root-job-id>
(alias --root) to pin the worker to one orchestration run: it then runs only
jobs whose root_job_id matches that value plus the root coordinator job
itself. To restart the daemon without losing its persisted Claude token, use
gitmoot daemon restart (not stop + start); gitmoot daemon stop --forget-runtime-auth deletes the persisted token file.

Use gitmoot agent prompt <agent-or-template> when the user wants to reuse a
Gitmoot agent prompt in the current chat. Use gitmoot agent run for
coordinator delegation through a registered Gitmoot agent; Gitmoot will route to
ask, review, or implement and own worktrees, branch locks, commits, pushes, PRs,
and workflow advancement. To orchestrate an orchestra of agents,
gitmoot orchestrate <agent> "..." [--repo R] is sugar for gitmoot agent run <agent> --background "...". Use gitmoot agent ask only for analysis, planning,
or questions. Add --background only when the user wants a queued background
job. This is the same agent registry and runtime adapter path used by PR-comment
jobs; the plugin only helps the runtime discover this skill and does not replace
the gitmoot CLI.

Use gitmoot report bug --job <job-id> --preview when a failed, blocked, or
cancelled Gitmoot job needs a user-shareable bug report. Preview first by
default. Create with --create --yes only when the user explicitly asks or the
active workflow policy permits filing reports, then report the created or
existing GitHub issue URL back to the user.

SkillOpt train generation is durable and idempotent on resume. Each review
item's artifacts, item row, and options commit in one transaction the moment
that item finishes, so an interrupted generate phase loses only the item that
was in flight. Re-running gitmoot skillopt train continue regenerates ONLY the
items that are not yet complete; fully-generated items are skipped, so no
duplicate options are produced and completed work is never rewritten. If a single
item has some but not all of its options/artifacts persisted, resume returns a
hard error (item <id> has partial generated options; inspect or clear review options before continuing) rather than guessing.

Use gitmoot skillopt train recover --session <id> [--out-root path] [--json]
to recover the OPTIMIZER phase only. It re-imports or repairs the optimizer
candidate package and classifies the iteration (for example
already_completed_candidate, already_completed_no_candidate,
optimizer_active, or corrupted_unrecoverable). It does NOT release the
generation-phase lock or rebuild generation options; use train continue for
generation resume.

PR Comment Commands

Use GitHub PR comments as the public audit trail:

/gitmoot help
/gitmoot status
/gitmoot <agent> review [instructions]
/gitmoot <agent> implement [instructions]
/gitmoot ask <agent> [question]
/gitmoot retry <job-id>
/gitmoot cancel <job-id>
/gitmoot merge
/gitmoot resume <job-id> retry|continue|abort|answer [instructions]
@<agent> ask|review|implement [instructions]

Template Agents

Install or refresh the built-in thermo review template:

gitmoot agent template update thermo-nuclear-code-quality-review
gitmoot agent start thermo-review \
  --runtime codex \
  --repo owner/repo \
  --template thermo-nuclear-code-quality-review \
  --start-daemon

Coordinator recipes are built-in templates for the Orchestra pattern: a
coordinator that fans work out to ephemeral workers (no pre-registration) and
reconvenes them in one continuation. review-panel convenes a panel of
diverse-lens reviewers over a PR and synthesizes their verdict;
decompose-and-verify splits an implementation task into parallel file-disjoint
legs and runs a verify step that depends on all of them; verifier is the
minimal one-producer + independent-verify pair. Route any coordinator agent
through a recipe with the --recipe flag on gitmoot orchestrate:

gitmoot orchestrate project-planner "Review PR #123 in this repo." --repo owner/repo --recipe review-panel
gitmoot orchestrate project-planner "Implement the export feature described in the task." --repo owner/repo --recipe decompose-and-verify

Create a local custom prompt template:

mkdir -p agents
gitmoot agent template draft frontend-reviewer --output agents/frontend-reviewer.md
$EDITOR agents/frontend-reviewer.md
gitmoot agent template validate agents/frontend-reviewer.md
gitmoot agent template add frontend-reviewer --file agents/frontend-reviewer.md
gitmoot agent start frontend-reviewer \
  --runtime codex \
  --repo owner/repo \
  --template frontend-reviewer \
  --role reviewer \
  --capability ask \
  --capability review

After editing a local template file, refresh Gitmoot's cached snapshot explicitly:

gitmoot agent template diff frontend-reviewer
gitmoot agent template update frontend-reviewer

Template updates are versioned locally. gitmoot agent template show <id>
prints the current version, content hash, source commit, and promotion state.
Agents use the current promoted version by default, or a pinned version when
configured with a reference such as --template frontend-reviewer@v1.
Queued jobs keep the exact template content snapshot they were created with.

Discover templates by metadata:

gitmoot agent template list --runtime codex --output goal_file
gitmoot agent template list --tag review --capability ask
gitmoot agent template show frontend-reviewer

Draft and validate a captured template before installing it:

gitmoot agent template draft release-planner
gitmoot agent template validate .gitmoot/templates/release-planner.md
gitmoot agent template add release-planner --file .gitmoot/templates/release-planner.md
gitmoot agent prompt release-planner

Agent Job Contract

Every Gitmoot job should return a concise and truthful gitmoot_result JSON
object:

{
  "gitmoot_result": {
    "decision": "approved|changes_requested|blocked|implemented|failed",
    "summary": "Brief outcome.",
    "findings": [],
    "changes_made": [],
    "tests_run": [],
    "needs": [],
    "delegations": []
  }
}

Use blocked when work cannot continue without human input or external state.
Use failed when the attempted action errored. Do not report tests, changes, or
approvals that were not actually verified. Use delegations to request follow-up
work by named Gitmoot agents.

Safety Rules

Preserve existing behavior unless the job explicitly changes it. Keep work
scoped to the target repo. Do not commit generated data, caches, logs, secrets,
session archives, cloned helper repos, or large outputs unless explicitly
requested. Respect Gitmoot branch locks for implementation jobs.

Implementation jobs must not edit or push an implementation branch unless
Gitmoot assigned the job and the branch lock is held by the assigned agent.
Review and ask jobs should inspect and report without mutating branches unless
the task explicitly instructs otherwise.

Redact secrets from GitHub comments, job summaries, raw examples, and copied
command output.

When Unsure

Reread this SKILL.md, then inspect /gitmoot help, gitmoot status, and the
relevant job events before acting.