[FEATURE] PrePlanMode and PostPlanMode Hook Events

This issue has been inactive for 30 days. If the issue is still occurring, please comment to let us know. Otherwise, this issue will be automatically closed in 30 days for housekeeping purposes.

Please bring this hook. commenting here before this feature request goes stale

We need this!

Need this!

Bump!

There is an EnterPlanMode and ExitPlanMode hook - does that solve the problem? FYI, I have just dropped in a solution on a similar bug that gives you:

1. Session plan file naming convention can be anything (I use my project folder name)
2. Manages session plan name changes between Claude restarts, while consistently writing to a common plan location for the project
3. Stored in Project Repo
4. Built with Serena MCP usage (but not required, can target any folder - including team-shared folder locations)
5. Automatically "Takes out the trash" (deletes session plan files from previous sessions)

Hope it helps someone else like it helped me

+1 on this. My immediate use case is simple: I want a push notification (via ntfy.sh) when planning mode finishes so I can walk away and come back when it's ready for review.

Today there's no reliable way to do this — PostToolUse:ExitPlanMode doesn't fire, and the Stop hook is too broad since it fires on every response. A dedicated PostPlanMode event would solve this cleanly.

We had ExitPlanMode event in tooluse, so??? We better still have this.

Lifecycle hooks make total sense, but they're compensating for something that should be built-in — plans don't have a lifecycle at all right now. There's no status (draft vs approved vs done), no archiving on transition, no way for a new plan to know an old one existed. If plans had native identity and state transitions, hooks like Pre/PostPlanMode become clean extension points instead of the entire persistence mechanism. Wrote up the fuller proposal in #29445 — hooks + a plan registry would be the real combo.

I want to hook EnterPlanMode and cancel it and tell the agent to use a skill instead. Is this possible?

would use this immediately. we have a workflow where the agent writes implementation plans to a file before starting work, and being able to hook into plan mode transitions would let us auto-archive the previous plan.

right now we work around it with a PreToolUse hook that checks if the tool is Write and the target is a plan file. it's hacky and misses cases where the agent updates a plan in-place rather than rewriting it.

the PostPlanMode hook is the more valuable one for us - it would let us snapshot the plan to a versioned file (plan-v1.md, plan-v2.md) so we have a history of how the approach evolved during a session. currently that context is lost when the plan gets overwritten.

we run a bunch of automated workflows triggered by hooks already, here's an example of the kind of thing PrePlanMode/PostPlanMode would unlock - right now we work around it with custom skill entry points: https://github.com/m13v/social-autoposter/blob/main/skill/SKILL.md

While dedicated PrePlanMode/PostPlanMode events don't exist yet, you can hook into plan mode transitions today using existing matchers:
PostToolUse + ExitPlanMode — fires when plan mode exits:

{
  "hooks": {
    "PostToolUse": [{
      "matcher": "ExitPlanMode",
      "hooks": [{"type": "command", "command": "bash .claude/hooks/post-plan.sh"}]
    }]
  }
}

PermissionRequest + ExitPlanMode — fires before exit, can block:

{
  "hooks": {
    "PermissionRequest": [{
      "matcher": "ExitPlanMode",
      "hooks": [{"type": "command", "command": "bash .claude/hooks/on-plan-exit.sh"}]
    }]
  }
}

Exit code 0 = approve, 2 = block (with reason in stdout sent back to Claude).
Example: archive plan on exit + notify

PLAN_FILE=$(ls -t "${HOME}/.claude/plans"/*.md 2>/dev/null | head -1)
if [[ -n "$PLAN_FILE" && -f "$PLAN_FILE" ]]; then
    ARCHIVE="${CLAUDE_PROJECT_DIR:-.}/.plans/archive"
    mkdir -p "$ARCHIVE"
    cp "$PLAN_FILE" "$ARCHIVE/$(basename "$PLAN_FILE" .md)-$(date +%Y%m%d-%H%M%S).md"
fi
exit 0

Limitations:

• PreToolUse + EnterPlanMode is inconsistent across versions (#21282)
• PostToolUse + ExitPlanMode does not fire when the user accepts the plan with "clear context" (#20397)
• Plan content isn't in the hook input schema — read the plan file directly from ~/.claude/plans/
• No way to distinguish plan creation vs. plan update

For EnterPlanMode interception, a UserPromptSubmit hook that checks for /plan in the prompt is a rough fallback, though it won't catch programmatic entry.

+1 on this

+1 on this, this is a great improvement in productivity