agents-md-verification-map
Skill1 starsChanged 53 days agoMITView on GitHub
Use when auditing, creating, or tightening a project AGENTS.md verification map with concrete T0-T3 checks, repo-specific commands, anti-weakening rules, pending_external boundaries, or global-vs-project instruction compression.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- Fable-Forge/agents-md-verification-mapword for word, first seen Sep 3, 2026
Other versions1 project uses a version of this.
Word for word · 1
- Fable-Forge/agents-md-verification-map★ 0100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository20 other files it gives its agents.
- llms.txtllms.txt
- agent-skill-migrationSkill
- cocos-creator-mobile-gameSkill
- game-entity-movement-presentationSkill
- game-event-directorSkill
- game-ui-panelsSkill
- godot-narrative-workflowSkill
- narrative-game-prototype-plannerSkill
- obsidian-article-writebackSkill
- offline-ux-kitSkill
- product-image-prompt-iterationSkill
- product-teardownSkill
- reporting-disciplineSkill
- scene-anchor-systemSkill
- screen2layersSkill
- sol-luna-local-parallelSkill
- sprite-atlas-slicerSkill
- unfinished-work-auditSkill
- windows-ai-cli-recoverySkill
- xiaohongshu-copywriterSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. About 0 are common, 11 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈11
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 10
Write short, project-specific AGENTS.md guidance that maps change risk to real validation commands.
- Very specific: line 20
Read package.json, test folders, tool scripts, docs, project files, and existing QA notes only as needed.
- Distinctive: line 21
Use only commands and files that actually exist.
- Distinctive: line 25
Write the smallest useful map.
- Distinctive: line 30
Add one project-specific anti-weakening rule: do not delete, skip, loosen, fake, bypass, or relabel tests and acceptance gates to make an …
- Distinctive: line 32
Avoid weak long-lived wording.
- Very specific: line 33
Do not use placeholders such as <相关测试> or vague lines like “run relevant checks”.
- Distinctive: line 34
Do not invent scripts, paths, scenes, devices, credentials, services, or QA proof.
- Distinctive: line 35
Do not make every edit run the full suite.
- Distinctive: line 36
Do not claim local simulation is real external integration.
- Distinctive: line 38
Verify the policy edit.
- Distinctive: line 40
Run git diff --check at repo scope for T0 hygiene.
- Distinctive: line 41
If the file is untracked, manually inspect the full file because git diff may not show it.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
AGENTS.md Verification Map
Overview
Write short, project-specific AGENTS.md guidance that maps change risk to real validation commands. Keep global rules compact; put checkout-specific commands, source-of-truth paths, and acceptance boundaries in the project file.
Workflow
- Confirm target and authority.
- Identify the exact repo root and the
AGENTS.mdfile to edit. - Respect nested repo boundaries. If the user names a nested root, keep parent files read-only unless explicitly allowed.
- Preserve existing dirty or untracked work. Do not reset, stash, reformat, or widen the write set.
- Identify the exact repo root and the
- Audit real verification surfaces before writing.
- Read
package.json, test folders, tool scripts, docs, project files, and existing QA notes only as needed. - Use only commands and files that actually exist.
- For Godot/Cocos projects, distinguish static contracts, headless/import checks, editor preview, real-window/device checks, and external human review.
- For web projects, distinguish typecheck/lint/unit tests, build, browser/E2E, local simulation, real integration, and deployment.
- Read
- Write the smallest useful map.
- T0: doc/copy/config-only changes use repo-wide
git diff --checkplus manual current Diff inspection. - T1: nearest focused test, static check, or targeted validator.
- T2: fuller automated build, contract, smoke, headless, or E2E coverage for new behavior or cross-file changes.
- T3: release, payment, auth, safety, permissions, real platform, device, real-window, or external-human validation. Keep this separate from automation and mark unavailable evidence as
pending_external. - Add one project-specific anti-weakening rule: do not delete, skip, loosen, fake, bypass, or relabel tests and acceptance gates to make an implementation pass.
- T0: doc/copy/config-only changes use repo-wide
- Avoid weak long-lived wording.
- Do not use placeholders such as
<相关测试>or vague lines like “run relevant checks”. - Do not invent scripts, paths, scenes, devices, credentials, services, or QA proof.
- Do not make every edit run the full suite; map validation to risk.
- Do not claim local simulation is real external integration.
- Do not use placeholders such as
- Verify the policy edit.
- Re-read the edited block.
- Run
git diff --checkat repo scope for T0 hygiene. - If the file is untracked, manually inspect the full file because
git diffmay not show it. - Confirm the diff touches only the intended
AGENTS.mdunless broader writes were explicitly authorized.
Output Contract
Report the target file, evidence inspected, inserted or changed T0-T3 map, verification actually run, unverified T3/external gates, and any scope deliberately left untouched.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

