How do I install the Check skill?

Run npx skillmds@latest add christophecapel/check in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Check safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Check?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Check free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

christophecapel (@christophecapel) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Check by christophecapel — Agent Skill

Name: Check
Author: christophecapel

Check

Session-close audit — session-scoped manifest, two-section open items, decisive close signal. Runs the `/check` audit defined in `check.md`. Use when the user says "check", "audit this session", "close the session", "definition of done", or wants a clean-vs-blocked verdict before ending a session.

christophecapel 0530b57 2 files · 10.2 KB Updated Sep 17, 2026

File contents

/check — Session-Close Audit

This skill runs the session-close audit defined in check.md.

What it does

Builds a Session Manifest from chat-context (PRs, commits, tickets, memory files, edited files) — derives scope FIRST, before any audit
Walks 5 audit sections (Learnings, Changes, Docs, Issues, Memory) with a per-section emoji status (✅ / ⚠️ / ❌) reflecting only Section A items
Routes open items into two sections:
- Section A — In-session (blocks close if any row is tackle now)
- Section B — Adjacent context (never blocks close, informational only)
Pre-render check: validates PR-merge dependencies before flagging rows as blocked by user. If gh pr view shows the PR is already merged, the row auto-flips to tackle now and the dependent action gets executed in the same turn.
Emits a decisive verdict: ✅ Ready to close / ❌ Not ready / ⚠️ Close with caveats. Re-running /check for the same state produces the same verdict.

Why session-scoped

Earlier /check versions surfaced every open item across all sessions, projects, and pre-existing tracker tickets. That made every close run feel incomplete and triggered multi-/check loops to walk through items that didn't belong to the session. Session-scoped audit runs in one pass — items from other sessions land in Section B and never block close.

Usage

Type /check at any point you want a clean-vs-blocked verdict on the current session. The skill follows check.md step by step.

Implements mechanisms

---name: checkdescription: Session-close audit — session-scoped manifest, two-section open items, decisive close signal. Runs the `/check` audit defined in `check.md`. Use when the user says "check", "audit this session", "close the session", "definition of done", or wants a clean-vs-blocked verdict before ending a session.---# /check — Session-Close AuditThis skill runs the session-close audit defined in [`check.md`](check.md).## What it does- Builds a **Session Manifest** from chat-context (PRs, commits, tickets, memory files, edited files) — derives scope FIRST, before any audit- Walks 5 audit sections (Learnings, Changes, Docs, Issues, Memory) with a per-section emoji status (✅ / ⚠️ / ❌) reflecting **only Section A items**- Routes open items into two sections: - **Section A — In-session** (blocks close if any row is `tackle now`) - **Section B — Adjacent context** (never blocks close, informational only)- **Pre-render check:** validates PR-merge dependencies before flagging rows as `blocked by user`. If `gh pr view` shows the PR is already merged, the row auto-flips to `tackle now` and the dependent action gets executed in the same turn.- Emits a **decisive verdict**: ✅ Ready to close / ❌ Not ready / ⚠️ Close with caveats. Re-running `/check` for the same state produces the same verdict.## Why session-scopedEarlier `/check` versions surfaced every open item across all sessions, projects, and pre-existing tracker tickets. That made every close run feel incomplete and triggered multi-`/check` loops to walk through items that didn't belong to the session. Session-scoped audit runs in one pass — items from other sessions land in Section B and never block close.## UsageType `/check` at any point you want a clean-vs-blocked verdict on the current session. The skill follows `check.md` step by step.## Implements mechanisms- [#16 — Smallest shippable first](https://github.com/christophecapel/claude-mechanisms/blob/main/mechanisms/16-smallest-shippable-first.md)- [#11 — One branch, one scope](https://github.com/christophecapel/claude-mechanisms/blob/main/mechanisms/11-one-branch-one-scope.md)- [#1 — Discover and derive, never assume or ask](https://github.com/christophecapel/claude-mechanisms/blob/main/mechanisms/01-discover-and-derive.md)

Frequently asked questions