[Behavioral] Plan-Mode reminder persists after /plan-complete + /session-end
Behavioral Observation: Plan-Mode reminder persists after /plan-complete + /session-end
Category: Behavioral Pattern (not a bug or feature request)
Model: Claude Opus 4.6 (1M context) via Claude Code 2.1.132 (Claude Code)
Reproducibility: 100% — every session that runs /plan-complete followed by /session-end
Severity: low
Observed Behavior
After invoking /plan-complete to close a plan, the harness still keeps the active-plan-mode reminder injected. When /session-end is invoked immediately afterwards, it must call ExitPlanMode but the plan-context stub is gone — workaround: write a stub file before ExitPlanMode. Expected: /session-end (or any non-plan skill) implicitly resets plan-mode state, OR /plan-complete explicitly clears the reminder injection.
Expected Behavior
Plan-mode reminder cleared after /plan-complete, OR /session-end accepted as implicit ExitPlanMode confirmation
Claude's Self-Analysis
Stub-file workaround works but adds 1 extra Write + 1 Tool-Call per affected session, plus mental overhead in skill workflow
User Impact
Low: 1-2 tool-calls + 1 stub file per session that closes a plan and ends in same workflow. Cumulative across many sessions.
Local Mitigation
Stub-file workaround in /session-end skill body
Suggestion for Model Improvement
Either: (a) /plan-complete should clear plan-mode reminder injection upon successful completion, OR (b) /session-end should be whitelisted as implicit plan-mode-exit acknowledgement, OR (c) plan-mode-state should reset on any non-plan-related skill invocation that happens after the plan was marked complete
---
This observation was documented during a Claude Code session and reviewed by the user before submission.
No claims in this issue are estimated — all details are from actual session transcripts.
This issue has 1 comment on GitHub. Read the full discussion on GitHub ↗