agentleFS
andy-dendy-repoandy-dendy-repo/govern/.claude/skills/develop-feature/SKILL.md

develop-feature

Skill2 starsChanged 32 days agoMITView on GitHub

The development flow for any feature, use case, endpoint, entity or behaviour change in Govern: plan → implement layer by layer → unit/integration/architecture tests → XML docs → docs/ (RAG) → .claude rules correction → build + ALL tests green → code review → report. Use whenever the user asks to add, implement, build, change or fix functionality in src/.

3 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions3 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 3

    1. Koroqe/claude-code-sdlc★ 52same name
    2. a-novel-kit/stack★ 2same name
    3. hamburger-os/PraxFlow★ 1same 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 repository7 other files it gives its agents.

The file

Show its rules and text

We split this skill into the rules it gives an agent

15 rules. 0 are common, 2 are distinctive, and 13 are 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.

  • 13

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Follow the steps in order.

    line 12
  2. Very specific:

    Read .claude/rules/project-map.md and architecture.md, and look at existing features for patterns to reuse.

    line 15
  3. Distinctive:

    Ask the user (AskUserQuestion) about anything ambiguous that changes the design.

    line 16
  4. Very specific:

    Write the plan in the .claude/rules/planning.md format.

    line 17
  5. Very specific:

    Make sure you are on a feature branch from stage (git switch stage && git switch -c feature/<name>).

    line 18
  6. Very specific:

    Use cases go in [Feature]/Contracts/[UseCase]Contract.cs (Request / Response / Validator / Handler).

    line 25
  7. Very specific:

    Follow csharp-style.md and logging.md ([LoggerMessage], no sensitive data).

    line 27
  8. Very specific:

    Use roslyn-refactor tools for renames, moves, signature changes and extractions instead of manual multi-file edits.

    line 28
  9. Very specific:

    Every new public or protected member is documented (xml-docs.md).

    line 37
  10. Very specific:

    Update or create docs/features/<feature>.md and docs/README.md, plus docs/architecture.md, docs/operations/* or an ADR when relevant …

    line 40
  11. Very specific:

    If this work established or changed a convention, update the matching .claude/rules/*.md, skill or CLAUDE.md and the architecture test.

    line 43
  12. Very specific:

    Run /verify (build, format, arch, unit and integration tests, analyzers).

    line 46
  13. Very specific:

    Run the code-reviewer agent on the diff.

    line 49
  14. Distinctive:

    follow-ups and open questions

    line 56
  15. Very specific:

    Commit or push only when the user asks: conventional commit messages, on the feature branch, with the PR to stage.

    line 58

“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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Develop a feature

Follow the steps in order. Never report a feature as done without a green dotnet test Govern.slnx run in this session whose output you show. A Stop hook (.claude/hooks/require-tests.sh) also enforces this.

1. Understand and plan

  • Read .claude/rules/project-map.md and architecture.md, and look at existing features for patterns to reuse. Use cwm-roslyn-navigator (find_symbol, get_project_graph, get_endpoint_map) rather than reading whole files.
  • Ask the user (AskUserQuestion) about anything ambiguous that changes the design.
  • Write the plan in the .claude/rules/planning.md format. For non-trivial work, get approval before coding.
  • Make sure you are on a feature branch from stage (git switch stage && git switch -c feature/<name>). Never work on main or stage directly.

2. Implement layer by layer

Path-scoped rules load only when a matching file is read. Before creating files in a new area, read the matching rule files yourself: csharp-style, xml-docs and logging for src/, the test rules for tests/, and docs for docs/.

Domain → Application.Common → Application → Infrastructure → Api → Client. After each layer, run dotnet build Govern.slnx; warnings are errors.

  • Folders only as needed, per the templates. Names per the suffix table. No Helper classes.
  • Use cases go in [Feature]/Contracts/[UseCase]Contract.cs (Request / Response / Validator / Handler).
  • Schema change: dotnet ef migrations add <Name> --project src/Govern.Infrastructure --startup-project src/Govern.Api --output-dir Persistence/Migrations. Never hand-edit generated migration files.
  • Follow csharp-style.md and logging.md ([LoggerMessage], no sensitive data).
  • Use roslyn-refactor tools for renames, moves, signature changes and extractions instead of manual multi-file edits.

3. Tests

  • Unit (tests/Govern.UnitTests): only when Utils, Core or Domain changed (unit-tests.md).
  • Integration (tests/Govern.IntegrationTests): every new or changed endpoint needs the happy path, validation 400, auth 401/403 and domain errors (integration-tests.md).
  • Client (tests/Govern.ClientTests, bUnit): every new or changed client page, component, typed client or auth service (client-tests.md).
  • Architecture (tests/Govern.ArchitectureTests): if you introduced a new convention, folder or project, add or adjust a rule and prove it can fail (architecture-tests.md).

4. XML docs

Every new public or protected member is documented (xml-docs.md); the build fails on CS1591. Endpoints need <response> codes.

5. docs/ (RAG)

Update or create docs/features/<feature>.md and docs/README.md, plus docs/architecture.md, docs/operations/* or an ADR when relevant (docs.md). A game rules change also updates the player rules reference docs/game/rules.md and docs/game/game-logic.md (RulesReferenceTests checks the reference's tables against the ruleset).

6. Correct the Claude rules

If this work established or changed a convention, update the matching .claude/rules/*.md, skill or CLAUDE.md and the architecture test. Don't let rules drift from the code.

7. Verify

Run /verify (build, format, arch, unit and integration tests, analyzers). Loop fix → re-run until everything is green. If a test is flaky or can't run (e.g. Docker not running), say so explicitly; don't claim success.

8. Review

Run the code-reviewer agent on the diff. If auth, endpoints, CORS, configuration, infra or CI changed, also run the security-auditor agent. Fix Critical and High findings, then re-run step 7 if code changed.

9. Report

Keep it short:

  • what changed, as a list of files by layer
  • the test summary (total / failed / succeeded) copied from the real run
  • docs and rules updated
  • follow-ups and open questions

Commit or push only when the user asks: conventional commit messages, on the feature branch, with the PR to stage.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.