agentleFS
ww-w-aiww-w-ai/bkit-gemini/skills/pdca/SKILL.md

pdca

Skill66 starsChanged 7 months agoApache-2.0View on GitHub

Unified skill for managing the entire PDCA cycle. Supports Plan → Design → Do → Check → Act workflow with automatic phase progression.

7 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

19 rules. 0 are common, 8 are distinctive, and 11 are very specific to this project.

  • 0

    Common

    Most of the 7 edited versions we read keep these.

  • 8

    Distinctive

    This author's own choices. Worth a look.

  • 11

    Very specific

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

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

    Check if docs/01-plan/features/{feature}.plan.md exists

    line 80
  2. Distinctive:

    If not, create based on plan.template.md

    line 81
  3. Very specific:

    Context Anchor Generation: After generating Plan document, extract Context Anchor (WHY/WHO/RISK/SUCCESS/SCOPE) from Executive Summary, …

    line 83
  4. Very specific:

    Write as ## Context Anchor table between Executive Summary and Section 1.

    line 85
  5. Distinctive:

    Verify Plan document exists

    line 96
  6. Very specific:

    Create Task: [Design] {feature} (blockedBy: Plan)

    line 98
  7. Distinctive:

    Ask user which option to select before proceeding

    line 106
  8. Distinctive:

    Verify Design document exists

    line 109
  9. Very specific:

    Create Task: [Do] {feature} (blockedBy: Design)

    line 111
  10. Distinctive:

    Full Upstream Context Loading: Read ALL upstream documents:

    line 112
  11. Very specific:

    Parse --scope parameter: If --scope module-1 provided, filter implementation guide to matching modules only from Design's Session Guide

    line 124
  12. Distinctive:

    Ask user to confirm scope before starting implementation

    line 126
  13. Very specific:

    After gap-detector analysis, also call bkit_iterate MCP tool for quantitative gap measurement:

    line 133
  14. Very specific:

    If matchRate < 90%, recommend calling /pdca iterate

    line 136
  15. Distinctive:

    Check results (when matchRate < 90%)

    line 139
  16. Very specific:

    Call bkit_iterate MCP tool for automated gap analysis:

    line 143
  17. Very specific:

    Call bkit_iterate again to re-check

    line 147
  18. Very specific:

    When running QA, call bkit_qa_run MCP tool:

    line 152
  19. Distinctive:

    If passRate < 95%, fix issues and re-run

    line 157

“Common” is measured: kept by at least 6 in 10 of the 7 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

PDCA Skill

Unified Skill for managing PDCA cycle. Supports the entire Plan → Design → Do → Check → Act flow.

Arguments

ArgumentDescriptionExample
plan [feature]Create Plan document/pdca plan user-auth
design [feature]Create Design document/pdca design user-auth
do [feature]Do phase guide/pdca do user-auth
analyze [feature]Run Gap analysis (Check)/pdca analyze user-auth
iterate [feature]Auto improvement (Act)/pdca iterate user-auth
report [feature]Generate completion report/pdca report user-auth
archive [feature]Archive PDCA documents/pdca archive user-auth
statusShow current status/pdca status
nextGuide to next phase/pdca next

Action Details

plan (Plan Phase)

  1. Check if docs/01-plan/features/{feature}.plan.md exists
  2. If not, create based on plan.template.md
  3. Create Task: [Plan] {feature}
  4. Context Anchor Generation: After generating Plan document, extract Context Anchor (WHY/WHO/RISK/SUCCESS/SCOPE) from Executive Summary, Requirements, and Risk sections. Write as ## Context Anchor table between Executive Summary and Section 1.

    DimensionContent
    WHYCore problem being solved
    WHOTarget users/stakeholders
    RISKTop 3 risks identified
    SUCCESSMeasurable success criteria
    SCOPEIn-scope / out-of-scope boundary

design (Design Phase)

  1. Verify Plan document exists
  2. Create docs/02-design/features/{feature}.design.md
  3. Create Task: [Design] {feature} (blockedBy: Plan)
  4. Context Anchor Embed: Copy Plan's ## Context Anchor table to Design document top (between header metadata and ## 1. Overview). If Plan has no Context Anchor, skip gracefully.
  5. Generate 3 Architecture Options:
    • Option A — Minimal Changes: Least modification, maximum reuse
    • Option B — Clean Architecture: Best separation of concerns
    • Option C — Pragmatic Balance: Good boundaries without over-engineering
  6. Present comparison table with trade-offs (complexity, maintainability, effort, risk)
  7. Ask user which option to select before proceeding

do (Do Phase)

  1. Verify Design document exists
  2. Provide implementation guide
  3. Create Task: [Do] {feature} (blockedBy: Design)
  4. Full Upstream Context Loading: Read ALL upstream documents:
    • PRD (docs/00-pm/{feature}.prd.md) — WHY context
    • Plan (docs/01-plan/features/{feature}.plan.md) — Context Anchor, Success Criteria
    • Design document — architecture decisions
  5. Decision Record Chain Display:
    Decision Record Chain
    [PRD] Target: {segment} — {rationale}
    [Plan] Architecture: {option} — {rationale}
    [Design] Pattern: {approach} — {rationale}
  6. Display Context Anchor from Design document header
  7. Parse --scope parameter: If --scope module-1 provided, filter implementation guide to matching modules only from Design's Session Guide
  8. Ask user to confirm scope before starting implementation

analyze (Check Phase)

  1. Call gap-detector agent
  2. Compare Design vs implementation
  3. Calculate Match Rate
  4. Create Task: [Check] {feature}
  5. After gap-detector analysis, also call bkit_iterate MCP tool for quantitative gap measurement:
    • This tool compares design document vs actual implementation
    • Returns matchRate percentage and specific gaps with file references
    • If matchRate < 90%, recommend calling /pdca iterate

iterate (Act Phase)

  1. Check results (when matchRate < 90%)
  2. Call pdca-iterator agent
  3. Auto-fix and re-verify
  4. Max 5 iterations
  5. Call bkit_iterate MCP tool for automated gap analysis:
    • Input: feature name, project directory, target match rate (default 90%)
    • Tool returns gaps with specific file:line references
    • Fix the identified gaps
    • Call bkit_iterate again to re-check
    • Repeat until matchRate >= 90% or max 5 iterations
    • Server tracks iteration count internally

qa (QA Phase)

When running QA, call bkit_qa_run MCP tool:

  • Input: feature name, project directory, test levels [1,2,3,4,5]
  • The tool auto-detects test framework (jest/vitest/playwright)
  • L1: Unit tests, L2: API tests, L3-L5: E2E/browser tests (if playwright installed)
  • Review results: passRate, coverage, defects
  • If passRate < 95%, fix issues and re-run

report (Completion Report)

  1. Verify Check >= 90%
  2. Call report-generator agent
  3. Create completion report
  4. Load ALL upstream documents for comprehensive reporting:
    • PRD — compare original value proposition vs delivered value
    • Plan — compare planned Success Criteria vs actual results
    • Design — note architecture decisions and any deviations
    • Analysis — include final Match Rate
  5. Include Decision Record summary and Success Criteria final status

PDCA Flow

[Plan] ✅ → [Design] ✅ → [Do] ✅ → [Check] 🔄 → [Act] ⏳ → [Report] 📋
                                       ↑_________|
                                    (if < 90%)

References

  • ${extensionPath}/templates/plan.template.md
  • ${extensionPath}/templates/design.template.md
  • ${extensionPath}/templates/analysis.template.md
  • ${extensionPath}/templates/report.template.md

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.