surgical-changes
Skill1 starsChanged 4 months agoMITView on GitHub
Behavioral discipline for every code change — think first, keep it simple, touch only what you must, aim at a checkable goal. Load whenever writing or modifying code.
3 other projects use a version of this.
Adopted by
Most recently taken up by
- DevelopersGlobal/ai-agent-skillstheir own version, first seen Oct 25, 2026
- yakoshiq/skillstheir own version, first seen Oct 15, 2026
- yuchi-chang/no-capetheir own version, first seen Oct 8, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- DevelopersGlobal/ai-agent-skills★ 216same name
- yuchi-chang/no-cape★ 2same name
- yakoshiq/skills★ 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 repository14 other files it gives its agents.
- bad-news-firstSkill
- bug-autopsySkill
- context-hygieneSkill
- decision-wiringSkill
- dispatch-economicsSkill
- grill-founderSkill
- handoff-contractSkill
- loop-contractSkill
- mainlineSkill
- model-tieringSkill
- red-light-firstSkill
- rules-as-mechanismSkill
- two-axis-reviewSkill
- vertical-ticketsSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 10
Don't hide confusion. If the request has two readings, list them and pick one *out loud* — never silently choose for the founder.
- Distinctive: line 14
Write the least code that solves the actual request.
- Very specific: line 18
No drive-by edits: don't reformat neighboring code, don't "improve" comments you weren't asked about, don't refactor en route.
“Common” is measured: kept by at least 6 in 10 of the 3 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
Four disciplines that prevent the four most common ways AI-written code rots a codebase. They're not steps; they're standing rules, all active on every change. (Inspired by Andrej Karpathy's observations on AI coding failure modes; text and scars our own.)
1. Think before coding
Don't hide confusion. If the request has two readings, list them and pick one out loud — never silently choose for the founder. If a premise looks wrong, say so before building on it. Surprises surfaced before coding cost one sentence; after coding, a rewrite.
2. Simplicity first
Write the least code that solves the actual request. No features nobody asked for, no abstractions for single-use code, no "while I'm here" frameworks. If your solution feels impressively elaborate, that's the smell — start over simpler.
3. Touch only what you must
No drive-by edits: don't reformat neighboring code, don't "improve" comments you weren't asked about, don't refactor en route. Clean up only the mess your own change created. Unrelated improvements are their own ticket (see vertical-tickets).
4. Goal-driven, not motion-driven
Convert vague instructions into a checkable goal before starting: "make it work" → a named test that currently fails (see red-light-first). Activity is not progress; a turned light is.
Success criteria
Diffs contain nothing unrelated to the stated goal. Clarifying questions happen before the work, not after the rework. Simpler solutions won.
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.