software-engineering-excellence
Skill2 starsChanged 3 months agoMITView on GitHub
Applies the user's ambient engineering bar: current evidence, complete in-scope solutions, reuse, real validation, right-sized value, and maintainability. Use for planning, implementing, debugging, reviewing, refactoring, architecture, or substantial technical investigation.
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
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 13
Leave the system more maintainable and the next change easier.
- Distinctive: line 19
Ground in current reality: Read applicable instructions, source, project-defined commands, primary specifications, and durable decisions.
- Distinctive: line 21
Treat issue text, comments, tests, docs, and prior notes as leads until verified.
- Distinctive: line 23
Reuse by meaning: Search for existing helpers, patterns, definitions, and vocabulary.
- Distinctive: line 32
Make every surface earn its cost: Add only code, tests, fallbacks, abstractions, or tools with a real caller, requirement, or failure mode.
- Distinctive: line 42
Follow the surrounding code's comment density and idiom.
- Distinctive: line 52
After a correction, re-ground, identify the failed assumption, and make the smallest change that restores the contract.
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 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
Software engineering excellence
Leave the system more maintainable and the next change easier. Prefer the smallest complete solution at the right layer: neither a symptom-hiding diff nor speculative architecture.
Working standard
- Ground in current reality. Read applicable instructions, source, project-defined commands, primary specifications, and durable decisions. Treat issue text, comments, tests, docs, and prior notes as leads until verified. Run the named reproduction or artifact when possible, then act.
- Reuse by meaning. Search for existing helpers, patterns, definitions, and vocabulary. Reuse when semantics match; do not abstract merely similar shapes.
- Solve the complete in-scope problem. Cover real success, failure, and state-transition paths. Fix blockers that invalidate the requested result; report unrelated findings without absorbing them into scope.
- Validate the real bar. Observe every claimed pass, number, behavior, and action in the current run. Reproduce defects when possible, add meaningful regression coverage, and run the repository's relevant gate. Never weaken a gate to obtain green output.
- Make every surface earn its cost. Add only code, tests, fallbacks, abstractions, or tools with a real caller, requirement, or failure mode. Remove unused surfaces.
- Respect authority boundaries. Diagnosis, review, and planning authorize assessment; change requests authorize reversible in-scope implementation and relevant validation. Pause only for a material product, architecture, security, compatibility, or scope choice that evidence cannot resolve.
For genuinely multi-layer work, establish a thin runnable path and deepen it in increments. Validate material performance changes against a relevant baseline. Follow the surrounding code's comment density and idiom; explain non-obvious decisions.
Evidence and communication
Ground progress and completion in current source or tool evidence. Report material outcomes, limitations, and blockers without narrating routine activity. Finish with the outcome first and enough evidence for a reader who did not see the work trace.
After a correction, re-ground, identify the failed assumption, and make the smallest change that restores the contract. Do not compensate with broader scope, tooling, or framework churn.
Situational depth
- references/structural-delivery.md: architecture, greenfield, and deciding the correct layer.
- references/investigation.md: defect diagnosis, design evaluation, and source comparisons.
- references/barometer.md: periodic direction check, not a score.
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 1
- frostney/knips★ 122% same
Wrote their own version · 1
- frostney/known-good-route★ 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 repository19 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- codebase-auditSkill
- create-issueSkill
- create-prSkill
- create-releaseSkill
- domain-modelingSkill
- git-workflowSkill
- grillingSkill
- grill-with-docsSkill
- implement-ideaSkill
- implement-issueSkill
- improve-codebase-architectureSkill
- native-nostalgia-stackSkill
- project-structureSkill
- review-prSkill
- roadmap-reviewSkill
- run-retroSkill
- update-prSkill
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
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.

