Agent Skills: execute-plan

Execute a plan

UncategorizedID: rstacruz/agentic-toolkit/execute-plan

Install this agent skill to your local

pnpm dlx add-skill https://github.com/rstacruz/agentic-toolkit/tree/HEAD/skill/atk/execute-plan

Skill Files

Browse the full folder contents for execute-plan.

Download Skill

Loading file tree…

skill/atk/execute-plan/SKILL.md

Skill Metadata

Name
execute-plan
Description
"Execute a plan"
  1. Find the plan:

    • Find the plan file. It may be mentioned previously in the conversation, or ask the user if it can't be found.
    • Ensure that the plan file is on the disk as an .md file. If not, write it to artefacts/<title>.md first.
  2. Evaluate and clarify:

    • Evaluate if plan makes sense.
    • Check if the plan has structured Tickets:
      • If no tickets exist, use the question tool to ask the user:
        • Do they want to use $spec-implementation-plan to break work into tickets?
        • Or do they have feedback on the plan?
    • If plan has other ambiguities or open questions, ask the user for clarifications first.
  3. Prepare:

    • Create an empty progress file (artefacts/progress.md)
  4. Proceed:

    • Proceed with implementing the plan.
    • Do one ticket.
    • Before starting each ticket:
      • Read the plan file and progress file again. This ensures we're working with updated copies.
    • After finishing each ticket:
      • Ask a @general-opus agent to use $review-changes skill. Ask it to review the uncommitted changes.
      • Assess feedback. Address any P1 issues that makes sense to do.
      • Commit your work. Include the ticket ID in the commit message.
      • Document learnings:
        • Assess the conversation, summarise work done, include assumptions flagged
        • Identify potential roadblocks that future dev work might encounter (eg, errors, wrong decisions)
        • Append them to progress file (create if it doesn't exist) - this is to assist future work
  5. Post-implementation review, when done:

    • Ask @general-opus and @general-gpt-5-3-codex agents to use $review-changes skill. Ask it to review all changes, starting with the first commit.
    • Assess feedback. Address any P1 issues that makes sense to do.
    • If there was feedback, ask reviews again, then address again. Keep looping until there are no more changes to do.
    • Ask a @general-opus agent to use $refine-tests skill.
    • Apply recommendations from $refine-tests using @general-opus agent.

Important reminders:

  • Always ask for review after ticket - this greatly impacts build quality
  • Always re-read plan and progress files before every ticket - changes from external sources are expected