Back to list
benredmond

implement

by benredmond

0🍴 0📅 Jan 13, 2026

SKILL.md


name: implement description: Build and validate loop (BUILDER + validation) - implements the architecture, runs tests, iterates until passing. Writes code following the plan. argument-hint: [task-identifier]

Combines BUILDER (write code) and validation (run tests) in a tight loop.

You can find active tasks in ./.apex/tasks/ or run with: /apex:implement [identifier] Load task file and begin implementation.

Contract rules:

  • Implementation MUST satisfy all AC-* or explicitly document unmet criteria
  • If scope/ACs must change, append a <amendment ...> entry inside task-contract and bump its version

Unplanned patterns can be submitted as new_patterns in apex_reflect, but NOT as patterns_used (which updates trust scores).

Run: Syntax → Formatting → Type check → Unit tests → Integration tests → Coverage

Return: Validation report comparing predictions vs reality

IF only warnings: → Document for review phase → Proceed

IF all pass: → Proceed to write implementation section

<implementation>
<metadata>
  <timestamp>[ISO]</timestamp>
  <duration>[Time spent]</duration>
  <iterations>[Build-validate cycles]</iterations>
</metadata>

<files-modified>
  <file path="[path]">
    <changes>[Summary of what changed]</changes>
    <patterns-applied>
      <pattern id="PAT:X:Y">[How it was used]</pattern>
    </patterns-applied>
    <diff-summary>[Key additions/removals]</diff-summary>
  </file>
</files-modified>

<files-created>
  <file path="[path]">
    <purpose>[Why created]</purpose>
    <patterns-applied>[PAT:IDs]</patterns-applied>
    <test-file>[Corresponding test if any]</test-file>
  </file>
</files-created>

<validation-results>
  <syntax status="pass|fail">[Details]</syntax>
  <types status="pass|fail">[Details]</types>
  <tests status="pass|fail" passed="X" failed="Y" skipped="Z">[Details]</tests>
  <coverage>[Percentage if available]</coverage>
</validation-results>

<acceptance-criteria-status>
  <criterion id="AC-1" status="met|not-met">[Evidence or reason]</criterion>
</acceptance-criteria-status>

<patterns-used>
  <pattern id="PAT:X:Y" location="file:line" outcome="worked|tweaked|failed">
    [Notes on usage]
  </pattern>
</patterns-used>

<issues-encountered>
  <issue resolved="true|false">
    <description>[What happened]</description>
    <resolution>[How fixed, or why unresolved]</resolution>
  </issue>
</issues-encountered>

<deviations-from-plan>
  <deviation>
    <planned>[What plan said]</planned>
    <actual>[What we did instead]</actual>
    <reason>[Why deviation was necessary]</reason>
  </deviation>
</deviations-from-plan>

<reviewer-handoff>
  <summary>[What was built]</summary>
  <key-changes>[Most important modifications]</key-changes>
  <test-coverage>[What's tested]</test-coverage>
  <known-limitations>[Edge cases, TODOs]</known-limitations>
  <patterns-for-reflection>[Patterns to report in apex_reflect]</patterns-for-reflection>
</reviewer-handoff>

<next-steps>
Run `/apex:ship [identifier]` to review and finalize.
</next-steps>
</implementation>

// Record implementation summary as evidence apex_task_append_evidence(taskId, "learning", "Implementation complete: " + summary, { files_modified: files_modified.length, files_created: files_created.length, patterns_applied: patterns_used.length, tests_passed: validation_results.tests.passed, iterations: metadata.iterations } )

// Checkpoint for phase completion apex_task_checkpoint(taskId, "BUILDER_VALIDATOR: Implementation complete, ready for review", 0.85)

</mcp-calls>
</step>

</workflow>

<critical-requirements>

<pattern-fabrication-prevention>
YOU CANNOT FABRICATE PATTERNS.

Only claim patterns that exist in `<plan><patterns><applying>`.
In `<patterns-used>`, only list patterns from the plan.
Pattern IDs claimed here will be validated during `/apex:ship`.

VIOLATION: Claiming "PAT:NEW:THING" that was never in the plan
CONSEQUENCE: apex_reflect will reject, trust scores become meaningless
</pattern-fabrication-prevention>

<syntax-gate>
Before completing implementation:
- Run linting
- Check for common errors (double async, missing brackets)
- Fix ALL syntax errors before proceeding
- DO NOT transition to ship with syntax errors
</syntax-gate>

<contract-gate>
Before finishing:
- Confirm all AC-* are met, or explicitly mark any unmet criteria with reasons
- If contract scope/ACs changed, record an amendment with rationale and bump contract version
</contract-gate>

<spec-unclear-protocol>
If implementation reveals spec ambiguity:
1. Document the ambiguity
2. Ask user for clarification
3. If architectural change needed, note it for plan revision
4. Do NOT guess and implement wrong thing
</spec-unclear-protocol>

</critical-requirements>

<success-criteria>
- All implementation steps from plan completed
- All validation gates passed
- Full test suite passing
- No syntax errors
- Acceptance criteria status reported for all AC-*
- Patterns used are from plan only (Pattern Evidence Gate passed)
- Deviations documented with reasons
- Task file updated at ./.apex/tasks/[ID].md
- apex_task_checkpoint called at start, per-step, and end
- apex_task_append_evidence called for pattern usage
- apex_task_update called with db_role: BUILDER_VALIDATOR
</success-criteria>

<next-phase>
`/apex:ship [identifier]` - Review, document, and reflect
</next-phase>

</skill>

Score

Total Score

60/100

Based on repository quality metrics

SKILL.md

SKILL.mdファイルが含まれている

+20
LICENSE

ライセンスが設定されている

+10
説明文

100文字以上の説明がある

0/10
人気

GitHub Stars 100以上

0/15
最近の活動

3ヶ月以内に更新がある

0/10
フォーク

10回以上フォークされている

0/5
Issue管理

オープンIssueが50未満

+5
言語

プログラミング言語が設定されている

+5
タグ

1つ以上のタグが設定されている

0/5

Reviews

💬

Reviews coming soon