Menu
Create and maintain execution plan files in tmp/ to reduce chat context for multi-step tasks
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Basé sur la classification professionnelle SOC
Create strict progress tables with verified numbers only - no placeholders or estimates
Stop and ask for direction when blocked instead of inventing assumptions or placeholder data
Validate CV data against strict 2-page ATS-compliant schema with comprehensive field checks and size constraints. Use when validating CV JSON before PDF generation, checking field constraints, ensuring 2-page layout compatibility, or verifying ATS compliance. Trigger on: "validate CV", "check CV data", "verify schema", "is this CV valid", or before any PDF generation.
Generate 2-page ATS-compliant PDF CVs using WeasyPrint with multi-language support (EN/DE/PL) and visual regression testing. Use when generating final PDF, running visual tests, debugging layout issues, or iterating on CV template design. Trigger on: "generate PDF", "create CV PDF", "render PDF", "visual regression test", or after successful CV validation.
| name | execplan |
| description | Create and maintain execution plan files in tmp/ to reduce chat context for multi-step tasks |
| argument-hint | ["task-name"] |
| disable-model-invocation | true |
Create a living execution plan document that keeps the conversation thread short.
Invoke this skill when:
/execplan cv_validation
/execplan pdf_generation_fix
The skill creates tmp/$ARGUMENTS_execplan.md with a structured plan template.
The generated plan file includes:
# Goal
One sentence describing what we're building/fixing.
# Acceptance Criteria
Observable behavior with exact verification commands:
- [ ] `npm test -- cv-visual.spec.ts` passes
- [ ] PDF size < 2 pages
- [ ] Visual diff < 5% threshold
# Plan
1. [ ] Smallest next step
2. [ ] Second step
3. [ ] Third step
# Commands
Exact commands with working directory and expected output.
# Validation
How to verify success (tests, manual checks).
# Recovery
Idempotence notes, rollback commands, retry steps.
# Notes
Append-only log of decisions and rationale with timestamps.
Update as you go:
[x] when completed[!] if blockedLink in responses: Use markdown links: tmp/cv_validation_execplan.md
Summarize deltas: Don't paste the full plan into chat - only summarize what changed.
If the plan involves OpenAI/Claude integration:
OPENAI_API_KEY)See .claude/workflows/execplan-examples.md for complete workflow examples.