Agent-Skills.md

Agent Skills: Plan Mode UI (Claude Code)

Plan automation for Claude Code (hooks, scripts, auto-reload). Handles auto-resume after /clear, context threshold detection, and session-keyed isolation. Use when the user says "execute plan", "load plan", "start the plan", "run the plan", "reload", or wants to work from a previously created plan. IMPORTANT - Always update the plan file after completing tasks.

["plan-mode""workflow""automation""ralph-loop""context-management""claude-code"]
UncategorizedID: tianjianjiang/smith/smith-plan-claude

Repository

tianjianjiang/smith
tianjianjiang
1

Install this agent skill to your local

pnpm dlx add-skill https://github.com/tianjianjiang/smith/tree/HEAD/smith-plan-claude

Skill Files

Browse the full folder contents for smith-plan-claude.

Download Skill

Loading file tree…

smith-plan-claude/SKILL.md

Skill Metadata

Name
smith-plan-claude
Description
ExitPlanMode UI pattern

Plan Mode UI (Claude Code)

ExitPlanMode UI pattern for Claude Code. For context management, auto-resume, and hooks, see @smith-ctx-claude/SKILL.md.

Load if: Using ExitPlanMode in Claude Code plan mode Prerequisites: @smith-ctx-claude/SKILL.md

Explain Before ExitPlanMode

The approval modal shows the plan FILE and draws the user's eye away from your chat message -- in-message prose is easily missed. Always put the explanation (plain text; zh-Hant per user preference) in its OWN turn FIRST, then call ExitPlanMode in a LATER turn. Do not bury a substantial explanation in the same turn as the ExitPlanMode call. There is no "plan is obvious enough to skip" exception -- deterministically enforced by the exit-plan-mode-guard.mjs PreToolUse hook (smith-ctx-claude/scripts/exit-plan-mode-guard.mjs, README.md Hooks), which blocks the call unless a prior turn in the exchange was plain-text-only elaboration.

ExitPlanMode Rejection Handling

ExitPlanMode rejection has three scenarios -- handle each differently:

  1. Rejection WITH user feedback (normal revision flow):

    • Read the user's feedback from the rejection message
    • Revise the plan file based on their feedback
    • Send the revised explanation as its own plain-text turn first (the exit-plan-mode-guard hook requires fresh elaboration after ANY prior ExitPlanMode attempt, approved or rejected -- the original elaboration does not carry over)
    • Call ExitPlanMode again with the updated plan, in a later turn
    • Never call ExitPlanMode twice without making changes between calls

  2. Rejection WITHOUT feedback + "Re-entering Plan Mode" (silent redirect):

    • This is a known Claude Code issue where ExitPlanMode bounces back
    • Do NOT enter an infinite edit-and-retry loop
    • Use AskUserQuestion to tell the user: "ExitPlanMode was silently redirected. Please exit plan mode manually (Escape or Shift+Tab), then tell me to proceed."

  3. Session ends after ExitPlanMode (auto-accept and clear):

    • User clicked "auto-accept and clear context" -- session is killed
    • Preemptive flag from inject-plan.sh ensures plan auto-reload in next session
    • No agent action needed (agent doesn't see this -- session already ended)

Related

  • @smith-plan/SKILL.md - Portable plan tracking protocol
  • @smith-ctx-claude/SKILL.md - Context management, auto-resume, hooks, state
  • @smith-ralph/SKILL.md - Ralph Loop phase boundaries, resume
  • @smith-checkpoint/SKILL.md - Checkpoint (Serena + Basic-Memory)