verifying-changes
Skill0 starsChanged 50 days agoMITView on GitHub
Use when deciding what to build, run, or test after a change, and before claiming that work is complete, fixed, or passing.
5 other projects use a version of this.
Adopted by
Most recently taken up by
- PostHog/posthog-fosstheir own version, first seen Nov 1, 2026
- raalarcon9705/gated-dev-agentstheir own version, first seen Nov 1, 2026
- Mr-Quin/danmaku-anywheretheir own version, first seen Oct 8, 2026
- BuilderIO/agent-nativetheir own version, first seen Oct 8, 2026
- teispace/teitunneltheir own version, first seen Sep 27, 2026
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- BuilderIO/agent-native★ 7.6ksame name
- PostHog/posthog-foss★ 728same name
- Mr-Quin/danmaku-anywhere★ 550same name
- teispace/teitunnel★ 2same name
- raalarcon9705/gated-dev-agents★ 0same 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 repository20 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- templates/CLAUDE.mdCLAUDE.md
- capturing-decisionsSkill
- choosing-subagent-modelsSkill
- committing-changesSkill
- debugging-systematicallySkill
- delegating-workSkill
- designing-changesSkill
- exploring-a-codebaseSkill
- injecting-dependenciesSkill
- planning-changesSkill
- pruning-testsSkill
- reflecting-on-changesSkill
- resolving-conflictsSkill
- reviewing-codeSkill
- writing-commentsSkill
- writing-logsSkill
- writing-project-instructionsSkill
- writing-skillsSkill
- writing-testsSkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 5 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 5 edited versions we read keep these.
5
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 8
Evidence before assertions: Never say complete, fixed or passing without having run the command and read its output.
- Distinctive: line 13
Run the suites the change can actually reach.
- Distinctive: line 19
Run the full suite once at the end, as the final gate.
- Distinctive: line 26
Not verified: say that, never imply it.
- Distinctive: line 38
When relying on a cached or filtered run, name the gap.
“Common” is measured: kept by at least 6 in 10 of the 5 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
Verifying changes
Evidence before assertions. Never say complete, fixed or passing without having run the command and read its output.
Scope while iterating, sweep at the end
- Run the suites the change can actually reach.
- Too wide — a full sweep per small fix — slows the loop someone is watching.
- Too narrow — a compile-check when a suite covers the change, or one platform when the edit was in shared code — is how a regression ships.
- A compile-check alone is enough only when no test covers the change.
- A delete or rename reaches every suite that references the symbol, not just the diff.
- Run the full suite once at the end, as the final gate. Not per edit.
Report honestly
- Tests fail: say so, with the output — including failures you did not cause.
- A step skipped: say which and why.
- Done and verified: say it plainly, without hedging.
- Not verified: say that, never imply it.
A confident summary over an unrun command is the most expensive thing you can produce: it ends the review that would have caught the problem.
Read the result, not the log
- On failure, read the structured result — a result bundle, JUnit XML, a test summary — not
thousands of console lines. Grepping for
errorreliably finds the wrong ones. - Read the count, not just the status. A typo'd test filter runs zero tests and exits green; "0 failures" can mean 0 tests.
- Incremental builds keyed on source fingerprints can skip a resource-only change entirely.
- When relying on a cached or filtered run, name the gap.
Passing is not done
- A green suite proves the code does what its tests check, not what was asked.
- Re-read the brief, plan or ticket and check each requirement against the diff. Tests written in the same change share the author's reading, so they cannot catch a misreading.
Before claiming done
- The command was run in this session, after the last edit
- Its output was read, not assumed
- The test count is non-zero and matches expectation
- The full suite was swept once, at the end
- Each requirement in the brief, plan or ticket was checked against the diff
- Anything unverified is stated as unverified
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.

