agentleFS
Sign inStart free
typvtypv/nest-turbo-starter/.claude/skills/plan/SKILL.md

plan

Skill224 starsChanged 3 months agoMITView on GitHub

Plan implementations, design architectures, create technical roadmaps with detailed phases. Use for feature planning, system design, solution architecture, implementation strategy, phase documentation.

We split this skill into the rules it gives an agent

12 rules. 0 are common, 7 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 164 edited versions we read keep these.

  • 7

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Create detailed technical implementation plans through research, codebase analysis, solution design, and comprehensive documentation.

    line 10
  2. Very specific:

    If invoked with a task description, proceed with planning workflow.

    line 14
  3. Very specific:

    Default: --auto (analyze task complexity and auto-pick mode).

    line 27
  4. Very specific:

    Add --no-tasks to skip task hydration in any mode.

    line 37
  5. Distinctive:

    Always honoring YAGNI, KISS, and DRY principles.

    line 51
  6. Distinctive:

    DO NOT implement code - only create plans

    line 85
  7. Distinctive:

    Ensure self-contained plans with necessary context

    line 87
  8. Distinctive:

    Write plan.md + phase files (persistent layer)

    line 101
  9. Very specific:

    Check ## Plan Context injected by hooks:

    line 109
  10. Very specific:

    After creating plan: node .claude/scripts/set-active-plan.cjs {plan-dir} Reports: Active plans → plan-specific path.

    line 114
  11. Distinctive:

    DO NOT create plans or reports in USER directory.

    line 118
  12. Distinctive:

    ALWAYS create plans or reports in CURRENT WORKING PROJECT DIRECTORY.

    line 119

“Common” is measured: kept by at least 6 in 10 of the 164 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

Planning

Create detailed technical implementation plans through research, codebase analysis, solution design, and comprehensive documentation.

Default (No Arguments)

If invoked with a task description, proceed with planning workflow. If invoked WITHOUT arguments or with unclear intent, use AskUserQuestion to present available operations:

OperationDescription
(default)Create implementation plan for a task
archiveWrite journal entry & archive plans
auditAdversarial plan audit
validateCritical questions interview

Present as options via AskUserQuestion with header "Planning Operation", question "What would you like to do?".

Workflow Modes

Default: --auto (analyze task complexity and auto-pick mode).

FlagModeResearchAuditValidationCook Flag
--autoAuto-detectFollows modeFollows modeFollows modeFollows mode
--fastFastSkipSkipSkip--auto
--hardHard2 researchersYesOptional(none)
--parallelParallel2 researchersYesOptional--parallel
--twoTwo approaches2+ researchersAfter selectionAfter selection(none)

Add --no-tasks to skip task hydration in any mode.

Load: references/workflow-modes.md for auto-detection logic, per-mode workflows, context reminders.

When to Use

  • Planning new feature implementations
  • Architecting system designs
  • Evaluating technical approaches
  • Creating implementation roadmaps
  • Breaking down complex requirements

Core Responsibilities & Rules

Always honoring YAGNI, KISS, and DRY principles. Be honest, be brutal, straight to the point, and be concise.

1. Research & Analysis

Load: references/research-phase.md Skip if: Fast mode or provided with researcher reports

2. Codebase Understanding

Load: references/codebase-understanding.md Skip if: Provided with scout reports

3. Solution Design

Load: references/solution-design.md

4. Plan Creation & Organization

Load: references/plan-organization.md

5. Task Breakdown & Output Standards

Load: references/output-standards.md

Workflow Process

  1. Pre-Creation Check → Check Plan Context for active/suggested/none
  2. Mode Detection → Auto-detect or use explicit flag (see workflow-modes.md)
  3. Research Phase → Spawn researchers (skip in fast mode)
  4. Codebase Analysis → Read docs, scout if needed
  5. Plan Documentation → Write comprehensive plan via planner subagent
  6. Audit Review → Use Skill tool: Skill(skill: "plan", args: "audit {plan-path}") (hard/parallel/two modes)
  7. Post-Plan Validation → Use Skill tool: Skill(skill: "plan", args: "validate {plan-path}") (hard/parallel/two modes)
  8. Hydrate Tasks → Create Claude Tasks from phases (default on, --no-tasks to skip)
  9. Context Reminder → Output cook command with absolute path (MANDATORY)

Output Requirements

  • DO NOT implement code - only create plans
  • Respond with plan file path and summary
  • Ensure self-contained plans with necessary context
  • Include code snippets/pseudocode when clarifying
  • Fully respect the ./docs/development-rules.md file

Task Management

Plan files = persistent. Tasks = session-scoped. Hydration bridges the gap.

Default: Auto-hydrate tasks after plan files are written. Skip with --no-tasks. 3-Task Rule: <3 phases → skip task creation.

Load: references/task-management.md for hydration pattern, TaskCreate patterns, cook handoff protocol.

Hydration Workflow

  1. Write plan.md + phase files (persistent layer)
  2. TaskCreate per phase with addBlockedBy chain
  3. TaskCreate for critical/high-risk steps within phases
  4. Metadata: phase, priority, effort, planDir, phaseFile
  5. Cook picks up via TaskList (same session) or re-hydrates (new session)

Active Plan State

Check ## Plan Context injected by hooks:

  • "Plan: {path}" → Active plan. Ask "Continue? [Y/n]"
  • "Suggested: {path}" → Branch hint only. Ask if activate or create new.
  • "Plan: none" → Create new using Plan dir: from ## Naming

After creating plan: node .claude/scripts/set-active-plan.cjs {plan-dir} Reports: Active plans → plan-specific path. Suggested → default path.

Important

DO NOT create plans or reports in USER directory. ALWAYS create plans or reports in CURRENT WORKING PROJECT DIRECTORY.

Subcommands

SubcommandReferencePurpose
/plan archivereferences/archive-workflow.mdArchive plans + write journal entries
/plan auditreferences/audit-workflow.mdAdversarial plan audit with hostile reviewers
/plan validatereferences/validate-workflow.mdValidate plan with critical questions interview

Quality Standards

  • Thorough and specific, consider long-term maintainability
  • Research thoroughly when uncertain
  • Address security and performance concerns
  • Detailed enough for junior developers
  • Validate against existing codebase patterns

Remember: Plan quality determines implementation success. Be comprehensive and consider all solution aspects.

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.

Check your setup, free