coding-rules
Skill39 starsChanged 9 days agoMITView on GitHub
Repository-aware implementation rules for minimal design surface, contract safety, representative patterns, and verifiable changes. Use when implementing, refactoring, or reviewing code.
4 other projects use a version of this.
Adopted by
Most recently taken up by
- shinpr/agentic-codetheir own version, first seen Nov 3, 2026
- bmcszk/agent-skillstheir own version, first seen Oct 13, 2026
- jam7/ai-dev-gatestheir own version, first seen Aug 23, 2026
- hukusuke1007/agent-skillstheir own version, first seen Apr 10, 2026
Other versions4 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 4
- shinpr/agentic-code★ 51same name
- hukusuke1007/agent-skills★ 8same name
- bmcszk/agent-skills★ 2same name
- jam7/ai-dev-gates★ 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 repository30 other files it gives its agents.
- ai-development-guideSkill
- documentation-criteriaSkill
- external-resource-contextSkill
- implementation-approachSkill
- integration-e2e-testingSkill
- llm-friendly-contextSkill
- recipe-add-integration-testsSkill
- recipe-buildSkill
- recipe-designSkill
- recipe-diagnoseSkill
- recipe-front-adjustSkill
- recipe-front-buildSkill
- recipe-front-designSkill
- recipe-front-planSkill
- recipe-front-reviewSkill
- recipe-fullstack-buildSkill
- recipe-fullstack-implementSkill
- recipe-implementSkill
- recipe-outcome-watchSkill
- recipe-planSkill
- recipe-prepare-implementationSkill
- recipe-quality-profileSkill
- recipe-reverse-engineerSkill
- recipe-reviewSkill
- recipe-update-docSkill
- requirement-convergenceSkill
- reviewee-judgmentSkill
- subagent-delegationSkill
- subagents-orchestration-guideSkill
- testingSkill
The file
We split this skill into the rules it gives an agent
11 rules. About 0 are common, 11 are distinctive, and 0 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.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 10
Read references/typescript.md only for TypeScript used in web frontend work, including React applications.
- Distinctive: line 33
When a change alters a boundary where external content or model output selects a tool's action, target, or destination, verify that those …
- Distinctive: line 35
Keep secrets and sensitive values out of source, client bundles, errors, and logs.
- Distinctive: line 36
Use parameterized data access and verify authorization at resource access points when applicable.
- Distinctive: line 41
Before adopting a pattern, API, or dependency:
- Distinctive: line 44
Check whether the pattern is representative where alternatives coexist.
- Distinctive: line 45
Follow the dominant compatible pattern or record why another existing pattern is required.
- Distinctive: line 52
When the changed behavior implements a runtime rule that exists in multiple locations, inspect the other instances for synchronization …
- Distinctive: line 56
Keep the change within the accepted outcome and responsibility boundary.
- Distinctive: line 57
Use names and structure that expose domain intent.
- Distinctive: line 58
Remove unused code and obsolete comments in the changed scope.
≈ “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
Coding Rules
Reference
Read references/typescript.md only for TypeScript used in web frontend work, including React applications. It does not apply to backend or non-web TypeScript. Read references/security-checks.md when the change crosses an input, authorization, secret, persistence, or output boundary.
Source Order
Follow, in order:
- User-required outcome, explicit constraints, repository rules, and actual consumer contracts
- Task, acceptance criteria, Design Doc, ADR, and Work Plan as the current implementation plan
- Representative repository patterns
- Language/framework defaults
Use a sourced project contract in preference to generic best practice. When a planned mechanism is unnecessary or incorrect, preserve user outcomes and explicit constraints while reducing it: document passage and a generated AC do not make that mechanism mandatory. Keep the affected plan and proof aligned with the selected reduction.
Minimal Design Surface
Deliver the current requirement with the least justified total complexity. Before detailing new state, fields, modes, abstractions, or component splits, establish what required result a direct implementation cannot deliver. Evidence of a possible benefit makes the addition a candidate; retain it only when necessary now and worth its total lifecycle cost over subtraction or reuse.
Private local implementation details and test fixtures are not new design surface. If an element matches both categories, treat it as design surface. If several sufficient options remain, prefer the one with lower lifecycle cost.
Contract and Boundary Safety
- Preserve contracts required by remaining consumers and explicit user constraints; authorized removal includes the retired entry point and its unused implementation. Absence of repository call sites does not establish absence of external consumers. Resolve uncertainty only when it changes the removal or compatibility decision.
- Validate untrusted input at the boundary and encode output for its destination. Untrusted input includes request data, external service responses, model or tool output, and stored data whose writer is untrusted or whose consumer needs a guarantee the store does not make.
- When a change alters a boundary where external content or model output selects a tool's action, target, or destination, verify that those values cannot exceed the operation scope and access rights already granted to the caller.
- Propagate or handle errors with useful context and keep failures observable.
- Keep secrets and sensitive values out of source, client bundles, errors, and logs.
- Use parameterized data access and verify authorization at resource access points when applicable.
- For persistent or shared state, verify when partial, stale, committed, and rollback-only states become observable.
Repository-Local Choice
Before adopting a pattern, API, or dependency:
- Inspect the changed feature and relevant siblings.
- Check whether the pattern is representative where alternatives coexist.
- Follow the dominant compatible pattern or record why another existing pattern is required.
- Escalate dependency/version or architecture choices when repository evidence cannot resolve them.
Nearby code is evidence, not authority by itself.
When choosing whether to reuse, extend, or add a data structure, check its semantic meaning, owning responsibility, lifecycle, and boundary or interoperability cost. Reuse or extend it when those dimensions remain compatible; choose a separate structure when sharing would merge responsibilities or lifecycles that can change independently.
When the changed behavior implements a runtime rule that exists in multiple locations, inspect the other instances for synchronization impact. Centralize or update them together when they share responsibility and contract; preserve separate implementations when their responsibilities can evolve independently.
Change Discipline
- Keep the change within the accepted outcome and responsibility boundary. Treat task target files as the expected set; include an adjacent file when the same outcome requires it and report the expanded write set.
- Use names and structure that expose domain intent.
- Remove unused code and obsolete comments in the changed scope.
- Optimize only from measurements or a sourced requirement.
- Refactor in reversible increments and run the focused verification after each behavior-affecting step.
- Report adjacent debt outside scope and expand the change only with an explicit in-scope reason.
Completion Gate
- Every addition is necessary for the required outcome and worth its cost over a smaller response
- Actual consumer contracts and explicit constraints remain satisfied
- Repository-local pattern choice is supported by evidence
- Errors, sensitive data, and persistent state boundaries are handled
- Focused and repository-required checks pass
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.

