Index of every page on this site: llms.txt
All skills
Run it yourself

Use when implementing a feature, adding an endpoint, or making a non-trivial code change that requires pre-flight validation, TDD cycle, scope enforcement, and a clean commit. Combines pre-flight + tdd + scope-check + quality-commit into one flow. Keywords: implement feature, add endpoint, TDD, new functionality, code change. Triggers on "implement feature" or "add endpoint".

Use this Skill: https://skilld.dev/gh/acedergren/agentic-tools/implement

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈97 tokens for metadata: the name and description. ≈1k when used: this file.

Description uses 4.9% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈390 characters, or 4.9%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Implement

When to Use

Load this skill when the user request matches the frontmatter description for Implement.

End-to-end feature implementation pipeline. Runs pre-flight validation, TDD cycle, scope enforcement, and quality commit as a single orchestrated flow.

Do NOT load this skill when the user only wants a focused subflow like /tdd or /quality-commit, or when the work is already finished and only needs verification or commit handling.

NEVER

  • Never skip pre-flight — writing code on a broken baseline wastes the whole TDD cycle.
  • Never add features not covered by tests during the Green phase — that's scope creep inside a test cycle.
  • Never use git add -A or git add . — stage specific files only.
  • Never proceed past bootstrap mock failure — it means your mock wiring is broken, not a Red/Green issue.
  • Never commit with pre-existing lint errors in files you didn't touch — note them in the commit message instead.
  • Never treat a passing test during Red as "good enough" — a test that passes before implementation isn't testing new behavior.

Abort Conditions

STOP the pipeline and ask the user if:

  • Pre-flight finds the workspace in a broken state (typecheck fails before your changes)
  • More than 5 files need modification (scope may be too large for one commit)
  • Bootstrap mock test fails after 2 attempts
  • Full suite regression caused by your changes

Decision Tree: Which Phase to Start From

Is the workspace clean and green?
├── No → Pre-flight: fix or ask before any code
└── Yes → Proceed

Does a test file already exist for the target module?
├── Yes → Run it first; confirm green baseline
└── No → Bootstrap mock is your first step

Does the task touch > 5 files?
├── Yes → Ask user to confirm scope before writing
└── No → Proceed

Pipeline

Phase 0: Pre-flight (< 30 seconds)

git status --short
npx tsc --noEmit 2>&1 | head -20

If monorepo: check if shared/library packages need rebuild (source newer than compiled output).

If any check fails: report and ask how to proceed before writing code.

Phase 1: Understand & Plan

Read target files. Identify module type (route handler, repository, plugin, utility, service, component). Check nearest test files for established mock patterns.

Brief summary: "I'll add X tests covering Y, then implement Z." Wait for confirmation if scope > 3 files.

Phase 2: Bootstrap Mock (1 test)

Write ONE minimal test that imports the module and verifies mocks resolve. Run it — must pass.

If it fails after 2 attempts: STOP. Mock wiring is broken (wrong module path, incorrect mock factory, missing type shim). Fix that before Red/Green.

Phase 3: Red — Write Failing Tests

Write tests for: happy path, edge cases, error cases.

Run the file — ALL new tests MUST fail. If any pass unexpectedly, the tests aren't testing new behavior.

Phase 4: Green — Minimum Implementation

Write the minimum code to make all tests pass. Then run — all tests MUST pass.

Do NOT add features not covered by tests. Do NOT optimize. Do NOT refactor existing code.

Phase 5: Scope Guard

git diff --name-only

Flag files that don't relate to the task:

  • Formatting-only changes → revert with git checkout -- <file>
  • Unrelated refactors → revert or split into separate commit
  • Docstring additions to untouched code → revert

Phase 6: Full Suite + Quality Gates

npx vitest run --reporter=dot
npx tsc --noEmit
npx eslint <changed files only>

Triage failures:

  • Test failure in your files → regression, fix it
  • Type error → fix it
  • Lint in your files → fix it
  • Lint in files you didn't touch → note in commit message, do not fix

Phase 7: Commit

git add <specific files>
git commit -m "type(scope): description

Co-Authored-By: Claude <noreply@anthropic.com>"

Arguments

  • $ARGUMENTS: What to implement
    • Example: /implement add rate limiting to POST /api/search
    • Example: /implement src/routes/admin/settings.ts — add PATCH endpoint for theme
    • If empty: ask the user what to implement

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts7mo5 checks · Risk SAFE
  • Gen Agent Trust Hub8mo

    The skill provides an automated feature implementation pipeline using standard developer tools such as git and vitest. It is rated as LOW severity due to a surface for indirect prompt injection, as it processes user-provided arguments and local file content without explicit sanitization or boundary markers.

  • Socket8mo

    No alerts

  • Snyk8mo

    Risk: LOW · No issues

  • Runlayer8mo

    1 file scanned · No issues

  • ZeroLeaks7mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 641a79b. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last month.

Steadyupdated 6 months ago

README badge

README badge for acedergren/agentic-tools/implement