role-developer
Skill0 starsChanged 4 months agoMITView on GitHub
Use this skill whenever you implement a feature, change production code, fix a bug, or refactor, including small changes, even if the user does not ask for a particular process. Provides a compact implementation workflow that routes to the relevant standards skills. Load standards-code and the language-specific standards skill alongside it.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- NikitaIT/claude-skillstheir own version, first seen Mar 9, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- NikitaIT/claude-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 repository34 other files it gives its agents.
- AGENTS.mdAGENTS.md
- mcpSkill
- pattern-orchestration-complexSkill
- pattern-task-breakdownSkill
- role-architectSkill
- role-code-reviewSkill
- role-data-engineerSkill
- role-orchestratorSkill
- role-product-engineerSkill
- role-qa-engineerSkill
- role-security-auditorSkill
- role-site-reliability-engineerSkill
- role-technical-writerSkill
- standards-analysisSkill
- standards-api-designSkill
- standards-codeSkill
- standards-databaseSkill
- standards-documentationSkill
- standards-go-concurrencySkill
- standards-go-performanceSkill
- standards-goSkill
- standards-go-testingSkill
- standards-markdownSkill
- standards-observabilitySkill
- standards-pulumi-typescriptSkill
- standards-pythonSkill
- standards-securitySkill
- standards-shellSkill
- standards-skill-authoringSkill
- standards-testingSkill
- standards-typescriptSkill
- tool-storeSkill
- workflow-debuggingSkill
- workflow-pi-extensionSkill
The file
We split this skill into the rules it gives an agent
12 rules. About 0 are common, 10 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈10
Distinctive · estimate
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
- Very specific: line 21
Always load standards-code before changing production code, plus the language skill for the files being edited.
- Very specific: line 37
When a task references stored specifications and the host provides store tools, load tool-store.
- Distinctive: line 47
Ask only blocking questions.
- Distinctive: line 53
Add a task list only for genuinely multi-step work.
- Distinctive: line 57
Keep the diff focused and avoid drive-by refactors.
- Distinctive: line 65
Prefer a failing regression test before a bug fix where practical.
- Distinctive: line 70
Run the narrowest relevant check first, then broaden if it passes.
- Distinctive: line 81
Do not start implementing when requirements or destructive actions are ambiguous.
- Distinctive: line 82
Do not mask errors or weaken tests to get a green result.
- Distinctive: line 83
Do not introduce a new dependency without a stated reason.
- Distinctive: line 84
Do not assume a delegated agent shares your loaded skills or store context.
- Distinctive: line 90
Before declaring the work done:
≈ “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
Provides: An implementation workflow, standards routing, guardrails, and a done-check for production code changes.
Quick Reference
Flow: Understand → Load standards → Plan the seam → Implement → Test → Verify → Report
Bias: smallest useful change, existing conventions, early verification
Never: mask errors to get green results, or start implementing while requirements are ambiguous
Load Standards
Always load standards-code before changing production code, plus the language skill for the files being edited.
Load these when relevant:
| Skill | When |
standards-testing | Tests, fixtures, fakes, or coverage changes |
standards-security | Auth, untrusted input, secrets, tenancy, or PII |
standards-api-design | Public or internal contracts |
standards-database | Schemas, queries, migrations, or transactions |
standards-observability | New logging, metrics, or tracing |
standards-documentation | Public APIs or user-facing docs |
workflow-debugging | A failure with no confirmed cause |
These skills are the maintained source, not this one. standards-code owns the language-neutral rules and the language skill owns syntax and naming, standards-security owns the controls, standards-testing owns the conventions including tiered coverage targets. Load them rather than implementing from memory of what this skill used to say, because they may have moved on.
When a task references stored specifications and the host provides store tools, load tool-store. If store tools are unavailable, keep requirements in the task list or the conversation and omit [store:id] references rather than inventing repository files to stand in for them.
Workflow
1. Understand
- Restate the requested behavior and the acceptance criteria.
- Inspect nearby code and project conventions before editing.
- Ask only blocking questions; otherwise state your assumptions and proceed.
2. Plan the seam
- Choose the smallest change location that achieves the behavior.
- Identify the verification command before writing code. If you cannot name it, you do not yet know what "done" means.
- Add a task list only for genuinely multi-step work.
3. Implement
- Keep the diff focused and avoid drive-by refactors.
- Preserve public contracts unless changing them is the task.
- Validate inputs at boundaries and keep errors actionable.
- When the task is a refactor, confirm tests already cover the current behavior before changing anything. If they do not, add them first: without them you cannot tell a refactor from a regression. Keep the suite green between increments rather than only at the end.
4. Test
- Work test-first where the behavior is well specified: write the failing test, make it pass, then clean up.
- Prefer a failing regression test before a bug fix where practical.
- Cover changed behavior, edge cases, and the error paths this change affects.
5. Verify
- Run the narrowest relevant check first, then broaden if it passes.
- Capture long output to a file and quote only the relevant lines.
6. Report
- Summarize what changed, which files were touched, and the verification results.
Guardrails
- Do not start implementing when requirements or destructive actions are ambiguous. Ask first.
- Do not mask errors or weaken tests to get a green result.
- Do not introduce a new dependency without a stated reason.
- Do not assume a delegated agent shares your loaded skills or store context. State both explicitly in the delegation.
Checklist
Before declaring the work done:
- The required standards were loaded for every domain touched
- The code follows local conventions in the area changed
- Relevant tests, build, and lint ran, or skipped checks are explained
- The diff contains no unrelated refactors or leftover debug code
- The final response lists changed files and verification evidence
Integration with Other Skills
With role-architect: decide structure before implementing across module boundaries.
With role-qa-engineer: expand test coverage beyond the changed behavior.
With role-code-review: self-review the diff before submitting.
With role-security-auditor: get an independent pass on security-critical code.
With workflow-debugging: establish cause before fixing an unexplained failure.
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.

