agentleFS
QuirkybrainQuirkybrain/claude-code-swe-framework/.claude/skills/implement-component/SKILL.md

implement-component

Skill1 starsChanged 53 days agoMITView on GitHub

Implement or modify one bounded Component using its validated Contract and constraints, then build, test, review as required, and pass the Component Quality Gate before STABLE. Use for feature work, bug fixes, and refactors with an explicit scope.

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 8 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    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
  1. Distinctive:

    Run when a Component/task is ready with sufficient Requirements, design, Contract, write scope, and Definition of Done.

    line 14
  2. Distinctive:

    Read the Contract, constraints, acceptance criteria, and existing implementation before editing.

    line 30
  3. Distinctive:

    Stop if implementation exposes a Requirement, architecture, or Contract gap.

    line 33
  4. Distinctive:

    Add or update focused tests and traceability required by the change.

    line 35
  5. Distinctive:

    If review causes edits, rebuild and rerun affected tests/review.

    line 39
  6. Distinctive:

    Do not alter Requirements, architecture, acceptance criteria, or Contract behavior without an approved upstream decision.

    line 45
  7. Distinctive:

    Every code addition, modification, bug fix, or refactor requires Build, Test, and Verify.

    line 46
  8. Distinctive:

    Test PASS alone does not satisfy DoD.

    line 47
  9. Very specific:

    Never integrate a non-STABLE Component.

    line 49
  10. Very specific:

    Stop at STABLE with current Gate evidence, at FAILED/UNVERIFIED with diagnostics, or when upstream clarification/redesign is required.

    line 70

≈ “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.

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

Implement Component

Purpose

Execute Lego-style Component development without changing upstream intent or declaring unverified code complete.

Trigger

Run when a Component/task is ready with sufficient Requirements, design, Contract, write scope, and Definition of Done.

Inputs

  • Task/Component ID, allowed write scope, current state, target revision, and DoD.
  • Validated Requirements, architecture/design, versioned Contract, Project Constraints, and acceptance criteria.
  • Existing code/tests, build commands, risks, Workflow Level, and required Gate/review.

Prerequisites

  • Required dependencies are VALID; upstream material decisions are resolved.
  • Parallel work has non-overlapping scopes, stable Contracts, correct base visibility, and an integration owner.
  • Set the Component to IMPLEMENTING; any changed STABLE Component becomes UNVERIFIED.

Procedure

  1. Read the Contract, constraints, acceptance criteria, and existing implementation before editing.
  2. Confirm task scope, base revision, workspace, dependencies, and validation commands.
  3. Reproduce a bug before fixing when feasible.
  4. Stop if implementation exposes a Requirement, architecture, or Contract gap.
  5. Implement the smallest cohesive change within the Component boundary.
  6. Add or update focused tests and traceability required by the change.
  7. Build or run the project's equivalent compile/static validation.
  8. Execute Adaptive Testing for the affected risk and regression scope.
  9. Perform required Quick/Code/Architecture/Security/Performance/Integration Review through an independent reviewer when needed.
  10. If review causes edits, rebuild and rerun affected tests/review.
  11. Evaluate the Component Quality Gate against current evidence.
  12. Mark STABLE only after Gate PASS; otherwise retain UNVERIFIED or set FAILED.

Decision Rules

  • Do not alter Requirements, architecture, acceptance criteria, or Contract behavior without an approved upstream decision.
  • Every code addition, modification, bug fix, or refactor requires Build, Test, and Verify.
  • Test PASS alone does not satisfy DoD.
  • The implementation author does not issue the final independent review verdict.
  • Never integrate a non-STABLE Component.

Outputs

  • Scoped code/configuration/test changes.
  • Changed paths and Component/Contract revisions.
  • Exact build/test/review commands and results, limitations, risks, and traceability updates.
  • Component state and Gate verdict.

Failure Cases

  • Build/test/review fails, expected behavior is ambiguous, or dependency is unavailable.
  • The required fix belongs to Contract, data, design, architecture, Requirement, test, or environment.
  • Workspace/base differs from delegated inputs.

Escalation

Capture evidence and use Failure Diagnosis before repeated patching. After one bounded correction cycle—or immediately for cross-layer evidence—use Failure Escalation. Route material upstream changes through Clarification/Architecture Decision and Change Impact.

Stop Conditions

Stop at STABLE with current Gate evidence, at FAILED/UNVERIFIED with diagnostics, or when upstream clarification/redesign is required. Do not proceed automatically into integration.

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.