cursor-maintain-verification-skill
Periodic pass that keeps a project's verification skill and feature map honest: parallel source readers per feature, one live session driving every feature, at most one PR of proven corrections. Use for /maintain-verification-skill or "audit the verify skill".
- License 许可
- MIT
- 最近更新
- 2026-11-14
- Stars
- ★ 0
- Forks
- 1
01 前言 Front Matter
- name
- cursor-maintain-verification-skill
- description
- Periodic pass that keeps a project's verification skill and feature map honest: parallel source readers per feature, one live session driving every feature, at most one PR of proven corrections. Use for /maintain-verification-skill or "audit the verify skill".
- disable-model-invocation
- true
02 资源文件 9
- 参考文档agents/openai.yaml43 B
- 许可证LICENSE.upstream1.0 KB
- 参考文档MIRROR.md1.2 KB
- 参考文档references/runtime/method.md6.2 KB
- 参考文档references/runtime/native-codex.md6.2 KB
- 参考文档references/runtime/original-codex.txt5.8 KB
- 参考文档references/runtime/runtime-files.json575 B
- 参考文档references/runtime/selection.md1.7 KB
- 参考文档references/runtime/t3.md5.0 KB
03 SKILL.md 正文
Runtime selection
Before runtime-dependent work, read execution selection. In a verified T3 host follow the T3 contract and its method reference; the inline section below is the native Codex branch, not literal T3 dispatch. Outside T3, native Codex follows that inline branch unchanged. If the host is ambiguous or a required capability is unavailable, report the gap; a permission refusal never authorizes switching branches.
Native Codex branch
Codex adaptation boundary
This mirror maintains only the explicitly selected project-local verification skill under .agents/skills/verify-*/. The coordinator owns the live pass and proof-orchestrator selects its proof level. Source readers may use native bounded delegation when exposed, but the sequential fallback is only a declared path here, not a live proof. A missing or unhealthy target is blocked with the attempted route and prerequisite. The maintenance flow never edits product code or global skills, and it never opens, publishes, or merges a PR without separate authorization. The bounded fixture reruns the real disposable CLI, performs syntax/version/health Doctor, records one source-wave result per feature, reconciles only its own feature files, cleans app state, persists before/after evidence, and reports FIXTURE_ONLY; a separate disposable web fixture was also re-driven through the native Codex in-app browser after drift removal with accessibility-tree and screenshot evidence. It does not claim live target coverage, PR, or native host activation.
Maintain a verification skill
A feature map rots the moment the app changes. This skill is the upkeep loop for a skill generated by /create-verification-skill (or any project-local verification skill with a feature map). The unit of rigor is the feature, not every sentence: cover every feature file from source and exercise every feature live, without terminalising every bullet.
Outcomes
Pick one, and say which:
- clean — every feature got source and live coverage; nothing worth shipping. No branch, no PR.
- changed — one PR ships proven doc, harness, or map corrections.
- blocked — coverage could not finish or a proven fix could not ship safely. Say exactly what blocked it.
Edit scope
Only edit the verification skill's own directory (its SKILL.md, features/, and any harness scripts it owns). Never edit product code during a run: a behavior the map describes that the app no longer does is either doc drift (fix the map) or a product regression (report it, don't paper over it in docs).
Pass
-
Locate the target. Find the verification skill to maintain: the project-local skill whose body has launch/drive sections and a feature map (usually
.agents/skills/verify-*/). Several candidates → ask which one; none → stop and point at/create-verification-skillinstead of inventing a target. -
Index hygiene. Read the feature map README and glob its sibling files. Fix missing, extra, duplicate, or dead entries. Lightweight; no generated inventory.
-
Source wave. One read-only subagent per feature file, launched concurrently. Each explains "how does this user-facing feature work?" from source, flags likely doc drift with citations, and returns one concise live-verification recipe. Children never drive the app and never edit files. Return shape: feature summary / source entry points / likely drift or none / one recipe.
-
Reconcile. Every feature file has a returned summary. Merge overlapping recipes into as few app states as practical. Spot-check cited drift; don't re-prove clean claims. Sweep recent churn for user-facing surfaces missing from the map — require a concrete source path before calling one missing.
-
Live pass. Required even when source looks clean. The coordinator owns all driving; follow the verification skill's own launch model — one long-lived instance driven serially for servers and UIs, or a fresh isolated session per drive for short-lived CLIs (the skill's Launch section decides, not this one). Exercise every feature at least once, and hold three invariants the whole pass, whatever the failure: (1) never drive an instance you haven't health-checked since it last did something surprising — doctor before first drive, doctor on each fresh session where sessions are the unit, doctor again after any failed drive, and where doctor can't see the failure (a wedged UI state on a healthy process), reset to a known state or relaunch rather than hoping; (2) evidence captured so far survives every cleanup, checked at its named location, not assumed; (3) nothing a drive started outlives that drive's usefulness — failed-iteration residue is cleaned whether the session is stuck, exited, or shared (for a shared instance, clean the residue, not the instance). A doctor failure caused by skill drift is drift: fix it under edit scope and retry once — restart whatever the fix invalidated, nothing more — before calling the pass
blocked. A feature that can't be reached isverified-unreachableonly with the concrete prerequisite (auth, entitlement, OS, external state) and the route attempted; if the map omits that prerequisite, that's drift. Any harness fix from triage gets re-driven live before it ships. Final teardown happens after the last drive of the run — including those re-proofs — so nothing outlives the run (evidence stays, per the skill). -
Triage. Wrong or missing user-POV description → doc drift, fix it. Working behavior the harness can't drive → harness gap, fix it; a harness fix follows the same helpers rule as generation (scripts executable, invocation documented in the skill body). App behavior that's actually broken → product gap; record it for the user, keep it out of this PR.
-
Ship or stop. For changed: one PR of proven corrections, re-read every changed file first. For clean or blocked: no PR, report the outcome and the coverage honestly.
Keep concise run notes (features covered, unreachable prerequisites, confirmed drift, outcome) in a scratch location; don't commit them.
04 同仓库更多 Skills
X MCP guide
ALWAYS read this before using any X tool (the X plugin, any X MCP, or Grok Bot's built-in public X reads) and again on any X error. X tools are included in Grok Bot at no cost: never estimate costs, check credits, or tell the user to buy credits. tools=0 / user-X-* not found is a setup failure, not a paywall. Covers Spaces, communities, and Lists (reads and list writes), X Chat (XChat, inbox, DMs, unread, message requests, Chat PIN, juicebox, xchat_lite: clone xchat-grokbot-helper yourself and decrypt locally), scheduled X calls (schedule_call, list_scheduled_calls and the other call tools), and livestreams (stream sources, broadcasts, scheduled broadcasts, broadcast chat, moderators, mutes). Chat, call, livestream, Spaces, or Lists tools missing while other X tools work: reconnect the X plugin.
X Money guide
Read this before the first X Money action in a session and again on any X Money error, refusal, or missing capability. Covers the approval rule for every action that moves money (always ask, every time, no exceptions), how the connection works (connection code plus passkey in the X app), how to guide the user through connect, reconnect, and revoke at x.com/i/money/settings/connections, and what each failure message means.
google-slides
Build and edit Google Slides through the google-slides MCP tools - slides, grids of cards and columns, text boxes, shapes, images, tables, styling, layout inspection, and per-slide PNG rendering. Use when creating a deck, laying out slide elements, matching a template, or verifying how a slide looks.
google-sheets
Build and edit Google Sheets with the google-sheets tools - values, formatting, charts, conditional formatting, structure, validation, named and protected ranges. Use when creating a spreadsheet or dashboard, formatting cells, adding or fixing charts, or reorganizing tabs and rows.
google-drive
Work with Google Drive through its MCP tools - search, read, create, organize, share, comment, and export files. Use when creating files in specific folders, moving or renaming, converting Markdown/CSV into Docs/Sheets, managing comments, or exporting a rendered PDF.
google-docs
Author and edit Google Docs with the google-docs tools - structured writing, character and paragraph styling, lists, tables, images, headers/footers. Use when creating a formatted document, editing document content, or applying styles.