methodology
Skill2 starsChanged 22 days agoMITView on GitHub
Install a discipline for building software with AI agents in which no agent certifies its own work: contracts with verify clauses, a separate evaluator, two state files. Use when the user says 'adopt the methodology', 'set up a contract', or wants PROJECT.md and log.md, or when they ask how agents should verify their work.
7 other projects use a version of this.
Adopted by
Most recently taken up by
- xiaoyangtx996/Searching-skilltheir own version, first seen Nov 8, 2026
- AsiaOstrich/EngramGraphtheir own version, first seen Nov 8, 2026
- AsiaOstrich/universal-dev-standardstheir own version, first seen Nov 1, 2026
- pavel-molyanov/molyanov-ai-devtheir own version, first seen Oct 25, 2026
- AlexHorovitz/skillstheir own version, first seen Oct 17, 2026
- stepanenkoviktor0110-boop/ai-dev-methodologytheir own version, first seen May 31, 2026
Used with
Projects with it also use refactor.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
GitHubSkillrefactor
Surgical code refactoring to improve maintainability without changing behavior. Covers extracting functions, renaming variables, breaking down god functions, improving type safety, eliminating code smells, and applying design patterns. Less drastic than repo-rebuilder; use for gradual improvements.
Also in 3 projects that use this
Other versions7 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 7
- pavel-molyanov/molyanov-ai-dev★ 299same name
- AsiaOstrich/universal-dev-standards★ 84same name
- 0x-Shashi/WEB3-AUDIT-SKILLS★ 61same name
- AsiaOstrich/EngramGraph★ 4same name
- stepanenkoviktor0110-boop/ai-dev-methodology★ 2same name
- AlexHorovitz/skills★ 1same name
- xiaoyangtx996/Searching-skill★ 1same name
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 repository38 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- arenaSkill
- codebase-oracleSkill
- geppettoSkill
- lean-subagent-devSkill
- overnight-runSkill
- parallel-worktreesSkill
- dependency-security-auditSkill
- dependency-upgradeSkill
- rule-to-hookSkill
- safe-mergeSkill
- agent-friendly-auditSkill
- architectSkill
- brainstormSkill
- engineering-principlesSkill
- grill-meSkill
- improve-codebase-architectureSkill
- sharpenSkill
- to-issuesSkill
- to-prdSkill
- handoffSkill
- instructions-auditSkill
- project-setupSkill
- retrospectiveSkill
- session-closeSkill
- which-skillSkill
- bug-fixSkill
- ci-cost-and-cadenceSkill
- evaluatorSkill
- flaky-test-huntSkill
- hillclimbSkill
- launch-readinessSkill
- migration-reviewSkill
- mutation-testSkill
- review-triageSkill
- skill-evalSkill
- tddSkill
- verify-appSkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 3 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
3
Distinctive
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 8
If you started this skill on your own, without the user typing its name or asking for what it does, first say in one or two lines what it …
- Very specific: line 17
New project, no code yet: follow ${CLAUDE_SKILL_DIR}/PROMPT-new-project.md.
- Very specific: line 18
Existing project that works, discipline for new work only: follow ${CLAUDE_SKILL_DIR}/PROMPT-adopt-forward-only.md.
- Distinctive: line 27
Ask which door applies if it is not obvious from the repository.
- Distinctive: line 50
Do not copy phase names, rubrics or tooling from the method's source project.
“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Methodology
If you started this skill on your own, without the user typing its name or asking for what it does, first say in one or two lines what it will do and what it will cost, and wait for a yes. A remark or a question that only matches its description is not asking for it. If the user asked for it, or the brief you are working from does, go ahead.
The method is in ${CLAUDE_SKILL_DIR}/METHODOLOGY.md. Read it before doing anything else; every
term below is defined there.
Pick the door
- New project, no code yet: follow
${CLAUDE_SKILL_DIR}/PROMPT-new-project.md. About an hour. - Existing project that works, discipline for new work only: follow
${CLAUDE_SKILL_DIR}/PROMPT-adopt-forward-only.md. One session, no audit of what exists. This is the right door for most existing projects. - After every increment, either way: have an evaluator grade it in a fresh context. If the
evaluatorskill is installed, run it. If not, start a new subagent or session as the evaluator in mechanisms 3 and 5 of METHODOLOGY.md: give it the contract and the code, never the conversation that built it, and have it grade each assertion with evidence it reproduced and mutation-verify every test it relies on. If the user adopts exactly one thing, it is this.
Ask which door applies if it is not obvious from the repository.
Templates
In ${CLAUDE_SKILL_DIR}/templates/:
| File | Use |
CLAUDE-rules.md | Operating-rules block to merge into the project's CLAUDE.md |
CLAUDE-rules-forward-only.md | The same for the forward-only door: process tiers, the trust line, the contagion rule |
PROJECT.md | Mutable state file: phase, open gate, next action, done / in-flight / blocked |
log.md | Append-only history, ## [YYYY-MM-DD] phase | event; a wrong entry is corrected by a new entry that names the one it supersedes |
contract.template.json | Annotated contract schema: assertions with verify clauses, pinnedValues, nonGoals, gradingRule |
Copy them into the project, fill every bracket, and do not keep the brackets' wording.
The three rules that carry most of the weight
- The builder never grades itself. Different context, hostile prior, no access to how or why it was built.
- Mutation-verify every guard. Introduce the defect, watch the test go red, revert. If the
mutation-testskill is installed, it does this with its assertions checked by a script. - Done means an adversary found nothing new, twice. Not "the task list is empty."
Do not copy phase names, rubrics or tooling from the method's source project; those are its clothes, not its skeleton. Multi-agent orchestration is an accelerant, not the method: all of this works with one model and two chat windows.
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.