本文へ移動
cccskills
無料GitHub で公開

gaia-steward-lane

Pick up one bounded Gaia Steward maintenance dispatch, do exactly it, and hand the result back for independent verification. Use when the user says "check the steward lane", "pick up maintenance work", "is there anything for me to fix", "run the steward lane", "/gaia-steward-lane", or when a scheduled routine wakes you with no other instruction. Also use before starting speculative repository cleanup: if Steward has not found the work, the work probably does not exist.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md6.0 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

gaia-steward-lane

The pickup point for Gaia Steward's bounded Class B lane.

Steward decides what maintenance exists and how much authority it carries. This skill is how you take one item of that work, stay inside its envelope, and hand it back in a shape that can be verified rather than trusted.

The one thing to understand first

You are not being asked to maintain this repository. You are being asked to resolve one finding, inside one envelope, and to stop.

Almost every failure mode of this skill is the same mistake: noticing something else worth fixing and fixing it. Don't. Note it in your report — Steward will observe it on its own next scan, and it will arrive with its own authority.

Getting work

gaia steward lane next --prompt

Three possible outcomes, all normal:

OutputMeaningWhat you do
A full prompt beginning # Tree Keeper dispatchThere is workFollow that prompt. It is the contract; this file is only how you got it.
Nothing to dispatch: no queued Class B debtSteward found nothingStop. Report that and end. A quiet lane is a healthy lane.
Nothing to dispatch: … maxInFlight ceilingWork is already outstandingStop. Someone else's dispatch is unverified. Do not start a second one.

Exit code 0 in all three cases. Idle is not failure.

If you want to see the state before asking for work:

gaia steward lane status

Doing the work

The rendered prompt carries everything that binds you: allowed paths, allowed commands, forbidden paths, stop conditions, and a proof contract. Read all of it before your first edit.

Two rules the prompt states and this file repeats because they are the ones most often broken:

  1. The allowed commands are your only mutation interface. If the envelope grants you no command for something, you describe the change — you do not reach around the missing command and do it by hand. This repository routes changes through tools for reasons those tools enforce and you cannot see: audit trails, validation, provenance.
  2. You may not widen your own authority. Needing a path, command, or decision you were not granted is a finding to report, not a permission to assume. Say so and stop; that is a successful dispatch.

A dispatch that concludes "the finding does not reproduce" is also successful. Do not manufacture a change to justify the run.

Handing it back

Capture two things: the diff, and a record of the commands that prove it.

git diff > candidate.diff

Write the proof transcript as steward-proof-transcript-v1, one or more entries per item in the packet's proof contract, numbered from 1 in the order the packet lists them:

{
  "schemaVersion": "steward-proof-transcript-v1",
  "entries": [
    {"contractIndex": 1, "command": "python scripts/validate.py", "exitCode": 0, "output": "…"}
  ]
}

Then verify:

gaia steward verify <debt-id> --diff candidate.diff --proof proof.json

Do not skip this and do not summarize the diff instead. Most of what a reviewer would check by eye is mechanical, and Steward checks it for free: scope against the envelope, proof coverage and exit codes, whether guards were weakened, whether unrelated debt appeared while you worked.

Exit codes: 1 reject, 3 escalate, 0 pending. 0 does not mean accepted. Steward has no authority to accept — it can only fail to object. If you get 0, say so plainly and hand the work to a human or a second reader; a run of --prompt will render the independent verifier's prompt for exactly that.

If you get 1 or 3, read the reasons. They are specific, and they are usually right. A reject on scope means you edited something you were not granted; fix that before anything else.

What you must never do here

  • Merge, push to main, tag, or release.
  • Widen founder/steward/POLICY.yaml, or edit anything under .gaia/steward/. The policy defines your authority and the receipts are the audit trail; an agent editing either is the failure this whole system exists to prevent.
  • Record an accept verdict for your own work. gaia steward lane record --verdict accept exists for a second reader, and its --note records who said so.
  • Take a second dispatch before the first is verified.

Optional External Operator Advisory (Outside Runtime Only)

Gaia Steward maintenance dispatches run with a strict zero-model budget. Model execution is strictly forbidden inside the automated Class A/B dispatch loop. However, an external operator inspecting maintenance debt outside the runtime may optionally run the read-only advisory runner (docs/agents/jev.md):

# Offline dry-run by default ($0 spend); add --live to query TypeSafe API
gaia steward scan --json > /tmp/debt.json
python scripts/jev_advisory.py --mode steward --input /tmp/debt.json

This inspection provides non-binding hints only, never reorders maintenance debt, never expands the dispatch envelope, and hands off to worker-luna when uncertain or offline.

When there is genuinely nothing

Say so in one line and stop. Steward is measured on its no-op rate as much as its closure rate — a system that always finds something to do is a system that manufactures work.

Where the rest of this lives

  • founder/steward/README.md — the policy in prose, and what each bound is for.
  • founder/steward/routines/ — the human contract for each routine, and cadences.
  • founder/STEWARD.md — why any of this is shaped the way it is.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

ci-churn

無料

Measure CI churn cost for a pull request: how many commits and how many CI-compute seconds were burned on avoidable "fix the CI" push rounds. Use when you want to understand the efficiency cost of a PR's iteration cycle, get actionable pre-push check suggestions, or include a churn summary in a pipeline report. Triggers: "ci churn", "how much CI did we burn", "avoidable commits", "ci efficiency", "wasted CI time", "churn report", "/ci-churn". Called automatically by /fp-drift at pipeline close.

日本語の概要は準備中です。原文の説明を表示しています。

gaia-research/gaia-skill-tree232026年10月13日 更新

Produce the rendered evidence a human-gated frontend PR owes its reviewer: capture each changed surface (plus the before state for anything modified, plus mobile width), then generate one self-contained local HTML page the founder can open and review without checking anything out. Use when a PR is gated by the "Frontend changes are HUMAN-GATED" section of CLAUDE.md, or when the user says "/design-gate", "screenshot evidence for this PR", "prep this for founder review", "build the design evidence page", "what does this look like before and after". The page comes from a template script, not hand-written markup. Output is session-local and never committed. Do NOT use for non-gated changes: typo/grammar/factual fixes in an unchanged block, dead link targets, alt/ARIA fixes with no visual change, cache-bust bumps, or refactors with no rendered diff.

日本語の概要は準備中です。原文の説明を表示しています。

gaia-research/gaia-skill-tree232026年10月13日 更新

One-fix-at-a-time, approval-gated design fix recovery. Given a checklist of fixes (handover doc, CHECKLIST.md, or inline list), applies each individually, presents a diff, and waits for an explicit APPROVE / REJECT / NO CHANGES decision before committing. Trigger phrases: /design-iteration, "go through the checklist one by one", "fix these one at a time", "bucket recovery", "design fix pass".

日本語の概要は準備中です。原文の説明を表示しています。

gaia-research/gaia-skill-tree232026年10月13日 更新

Playbook for correcting a named skill's star level to match its computed Trust Magnitude grade, with an independent post-calibration verification pass and a CLI-logged timeline event (never hand-written). Use when: recalibrating a skill's level, fixing a stale or drifted star rating, running a batch of `gaia dev calibrate` corrections, or confirming a past calibration actually landed. Triggers: "calibrate <skill>", "recalibrate the trust magnitude backlog", "promote/demote <skill> to N stars", "why is <skill> still N stars", "verify a calibration actually landed", "Hall of Heroes ranking is wrong".

日本語の概要は準備中です。原文の説明を表示しています。

gaia-research/gaia-skill-tree232026年10月13日 更新

Close out a completed EPIC / sprint end-to-end: author a fact-verified meta-post announcing what shipped, archive the EPIC's handovers, snapshot orchestrator memory, gate the integration→main merge on CI-green + explicit human go, drive the version bump (merge-commit conventions that auto-sync classifies), and verify the release reached the registry/PyPI. Use when the user says "close out EPIC <n>", "wrap up the sprint", "ship the meta shift", "/epic-close", or asks to finalize an integration branch that is ready to merge to main. Generalizes the EPIC 1002 (Yggdrasil II) close-out. This skill orchestrates — it authors founder docs directly (superadmin) and delegates code/regen to workers; it NEVER merges to main without the human gate.

日本語の概要は準備中です。原文の説明を表示しています。

gaia-research/gaia-skill-tree232026年10月13日 更新

Run this skill for Phase 3 of the evidence verification pipeline — the adversarial audit step. Use when you need to check the evidence data lake for bad data before ingestion: dead links, wrong URL formats (tree/ vs blob/), subjective wording ("elite", "high-quality"), stale migration notes, benchmark catalog misuse/vendor-claim leakage, or skills whose star evidence conflicts with classified evidence level. Triggers on phrases like: "audit the data lake", "adversarial check", "ev-adversarial-audit", "check for noise in evidence", "flag bad evidence", "run the audit phase", "quality check the by-type files", or any reference to Phase 3 of the pipeline.

日本語の概要は準備中です。原文の説明を表示しています。

gaia-research/gaia-skill-tree232026年10月13日 更新

gaia-research のスキルをすべて見る

このスキルの問題を報告する