agentleFS
Sign inStart free
cobusgreylingcobusgreyling/grok-build-arsenal/skills/plan-mode-orchestrator/SKILL.md

plan-mode-orchestrator

Skill6 starsChanged 5 months agoMITView on GitHub

For any non-trivial task, start in Plan Mode, produce a clear step-by-step plan with risks and verification method, and wait for explicit approval before any edits. Use when the user says "plan this", "use plan mode", "don't edit yet", or describes a complex feature/refactor/migration.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use subagents for broad investigation if the scope is large.

    line 24
  2. Very specific:

    Run grok inspect if configuration or skills may be relevant.

    line 25
  3. Distinctive:

    Do not proceed to implementation.

    line 38
  4. Distinctive:

    Ask for approval: "Does this plan look good?

    line 39
  5. Distinctive:

    Use other skills (tdd-intelligence, git-discipline, etc.) as appropriate.

    line 44
  6. Distinctive:

    All subsequent edits are traceable to the approved plan.

    line 51

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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 Mode Orchestrator

The single most important skill in this arsenal. Enforces disciplined agentic work.

When to Use

  • User asks for a plan or explicitly says "plan mode"
  • Task is larger than a single obvious edit
  • Risk of breaking changes, architecture impact, or many files
  • User says "don't write code yet"

Core Behavior

  1. Immediately enter or confirm Plan Mode mindset.
    • Explicitly state: "Entering Plan Mode. I will not edit files until you approve a plan."
  2. Investigate first (read-only where possible).
    • Use subagents for broad investigation if the scope is large.
    • Run grok inspect if configuration or skills may be relevant.
    • Gather evidence: relevant files, existing patterns, tests, risks.
  3. Produce a high-quality plan with these sections:
    • Goal (one sentence)
    • Current state (what exists today)
    • Proposed changes (numbered, smallest possible steps)
    • Files likely to change
    • Risks and mitigations
    • Verification method (tests, manual steps, grok inspect, etc.)
    • Estimated diff size / complexity
  4. Present the plan and STOP.
    • Do not proceed to implementation.
    • Ask for approval: "Does this plan look good? Any changes before I proceed?"
  5. After approval:
    • Execute one step at a time where possible.
    • Re-verify after each meaningful chunk.
    • Use other skills (tdd-intelligence, git-discipline, etc.) as appropriate.

Subagent Usage

For large investigations, spawn 2–4 subagents with clear charters (e.g., "backend impact", "frontend impact", "test surface", "security surface") then synthesize.

Success Criteria

  • User has explicitly approved the plan (or a revised version).
  • All subsequent edits are traceable to the approved plan.
  • Verification steps from the plan were actually performed.

Anti-Patterns to Avoid

  • Jumping into edits "just to explore".
  • Vague plans ("I'll refactor the auth layer").
  • Plans that are really just implementation notes.

Example Trigger Phrases

  • "Plan the migration to the new MCP protocol"
  • "Use Plan Mode to add the arena runner feature"
  • "Don't edit anything yet, just plan the security hardening"

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