agentleFS
ww-w-aiww-w-ai/bkit-codex/.agents/skills/plan-plus/SKILL.md

plan-plus

Skill69 starsChanged 8 months agoApache-2.0View on GitHub

Brainstorming-enhanced PDCA planning. 6-phase process with HARD GATE rule.

We split this skill into the rules it gives an agent

15 rules. 0 are common, 11 are distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 11

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use Plan Plus when:

    line 33
  2. Very specific:

    Call bkit_get_status to load current PDCA state

    line 43
  3. Distinctive:

    Read AGENTS.md for project context

    line 44
  4. Distinctive:

    Check git log for recent history (last 10 commits)

    line 45
  5. Very specific:

    Read bkit.config.json for project configuration

    line 47
  6. Distinctive:

    Ask ONE question at a time.

    line 53
  7. Distinctive:

    Ask exactly ONE question per turn

    line 62
  8. Distinctive:

    Wait for user response before proceeding

    line 63
  9. Distinctive:

    If user provides vague answer, ask a follow-up clarification

    line 65
  10. Distinctive:

    Best For: When this approach is optimal

    line 76
  11. Distinctive:

    Ask user to select preferred approach or combine elements.

    line 78
  12. Distinctive:

    Ask user to confirm the trimmed scope.

    line 89
  13. Distinctive:

    Ask: "Does this accurately capture your intent?"

    line 106
  14. Very specific:

    Use bkit_pdca_plan with the feature name to get template

    line 111
  15. Very specific:

    Call bkit_complete_phase(feature, 'plan') to record completion

    line 114

“Common” is measured: kept by at least 6 in 10 of the 3 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

Plan Plus - Brainstorming-Enhanced PDCA Planning

HARD GATE RULE (CRITICAL)

CRITICAL: Do NOT generate any code until ALL brainstorming phases are complete and the user has approved the final plan document. This is a HARD GATE - no exceptions. If the user asks for code during planning, respond: "We're still in the planning phase. Let's complete the plan first to ensure we build the right thing."

Overview

Plan Plus combines intent discovery from brainstorming methodology with bkit PDCA's structured planning. It produces higher-quality Plan documents by exploring user intent, comparing alternatives, and applying YAGNI review.

Use Plan Plus when:

  • Feature requirements are ambiguous or complex
  • Multiple implementation approaches exist
  • Stakeholder alignment is needed before implementation
  • The feature scope needs careful trimming

Process

Phase 0: Context Exploration (Automatic)

  1. Call bkit_get_status to load current PDCA state
  2. Read AGENTS.md for project context
  3. Check git log for recent history (last 10 commits)
  4. Scan docs/01-plan/ for existing plans
  5. Read bkit.config.json for project configuration

Output: Internal context understanding (not shown to user)

Phase 1: Intent Discovery

Ask ONE question at a time. Wait for user response before next question.

Questions (in order):

  1. "What is the core problem this feature solves?"
  2. "Who are the target users and their key needs?"
  3. "What does success look like? (measurable criteria)"
  4. "What constraints exist? (time, tech, team, budget)"

Rules:

  • Ask exactly ONE question per turn
  • Wait for user response before proceeding
  • Summarize each answer before asking the next
  • If user provides vague answer, ask a follow-up clarification

Phase 2: Alternatives Exploration

Based on Phase 1 answers, present 2-3 implementation approaches:

For each approach:

  • Summary: 1-2 sentence overview
  • Pros: 3-5 advantages
  • Cons: 3-5 disadvantages
  • Effort: Low / Medium / High / Very High
  • Best For: When this approach is optimal

Ask user to select preferred approach or combine elements.

Phase 3: YAGNI Review

Present a checklist of features from the selected approach:

  • Mark each as: [Must Have] / [Nice to Have] / [Won't Do]
  • Apply YAGNI principle: "You Aren't Gonna Need It"
  • Remove speculative features
  • Focus on MVP that delivers core value

Ask user to confirm the trimmed scope.

Phase 4: Incremental Design Validation

Present the plan document section by section for approval:

  1. User Intent (from Phase 1)
  2. Alternatives Explored (from Phase 2)
  3. YAGNI Review Results (from Phase 3)
  4. Scope Definition
  5. Requirements (Functional + Non-Functional)
  6. Success Criteria
  7. Risks and Mitigations
  8. Architecture Considerations

For each section:

  • Present draft content
  • Ask: "Does this accurately capture your intent?"
  • Revise if needed before proceeding

Phase 5: Plan Document Generation

  1. Use bkit_pdca_plan with the feature name to get template
  2. Fill template with validated content from Phases 1-4
  3. Write to docs/01-plan/features/{feature}.plan.md
  4. Call bkit_complete_phase(feature, 'plan') to record completion
  5. Present final document for user review

Output path: docs/01-plan/features/{feature}.plan.md

MCP Tool Integration

PhaseToolPurpose
0bkit_get_statusLoad PDCA state
0bkit_analyze_promptDetect intent
5bkit_pdca_planGet plan template
5bkit_complete_phaseRecord plan completion

References

For detailed process patterns, question templates, and examples, see: references/plan-plus-process.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.

Reviews can't be read right now.

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.