steward

A repository-specific guide for handling pull-request and continuous-integration events in the loopdive/js2wasm project.

In plain words
What is it for?
It helps respond to review or CI activity, avoid unsafe re-enqueuing, and follow the repository's documented merge process.
Why use it?
It helps agents judge whether a pull request is ready using the project's actual checks and merge-queue rules.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/loopdive/js2/steward
Any agent
npx skills add loopdive/js2 --skill steward
Clone the repo
git clone --depth 1 https://github.com/loopdive/js2

Made for: Claude Code, Codex.

Per session 92 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,725 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00092 $0.02725
Opus 5 $0.00046 $0.01362
Sonnet 5 $0.00018 $0.00545
Haiku 4.5 $0.00009 $0.00272

Measured 2d ago against content hash 2ee24476a077, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

steward scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/skills/steward/SKILL.md · 192 lines

How it starts

The opening of the file, as written. The whole thing — 192 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/steward — acting on PR events in this repo

Read this before acting on a CI or review event on a PR you opened or drive. It does not relax anything the generic rules state as "never" (no skipping or disabling tests; no rewriting someone else's history; no approving or merging). It states the repo conventions the generic defaults get wrong here, and each one below has bitten a real lane.

Authoritative sources this file summarises: docs/ci-policy.md and CLAUDE.md. Where they disagree with this file, they win — re-verify rather than trusting this summary's dates.

1. NEVER enqueue. The server enqueues.

.github/workflows/auto-enqueue.yml is the single enqueuer (#2786), triggered on required-check completion. A dev/agent must not enqueue, and must never re-enqueue: re-adding a PR that is in the in-flight merge group rebuilds that group and cancels its run. A re-enqueue loop on the queue head cost ~3.5h of cancellation churn on 2026-06-20.

Your job on a green PR is to make it enqueueable and confirm it was taken — not to push it into the queue.

2. Judge readiness from the real signals, not from mergeStateStatus

  • A SKIPPED required check SATISFIES branch protection. Jobs skipped by their own if:/path filter still publish a check run with conclusion skipped. On a docs-only PR equivalence-gate skips and the PR is still CLEAN. Do not decide readiness by counting six SUCCESS conclusions — you will not find them and will wrongly conclude the PR is not ready. Judge each REQUIRED check as green when its conclusion is success OR skipped.
  • Do not read mergeStateStatus as the enqueue gate at all. Since #4094 enqueueEligibility() in scripts/enqueue-green-prs.mjs decides from four real signals only — not draft, no hold label, mergeable === "MERGEABLE" (the field, not the status string), and every required check green. The function's own comment is explicit: "mergeStateStatus is deliberately NOT a parameter. It cannot be consulted even by accident." The old ENQUEUEABLE = {CLEAN, HAS_HOOKS} set is vestigial — read lines 631 and 1276, both of which say "Was !ENQUEUEABLE.has(...)".

Read the full file on GitHub · 192 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 192 lines · 0 tokens per session scan A 2ee24476a077

Subscribe to this mod's changes

steward is a skill published in the GitHub repository loopdive/js2 (59 stars, last pushed 2d ago), licensed Apache-2.0. It adds 92 tokens to every session and 2,725 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.