What problem does it solve? AI coding agents lose track of goals and progress when context windows fill up or sessions are cleared. This Skill stores task plans, findings, and progress logs as markdown files on disk, so multi-step work survives context compaction, /clear commands, and session restarts. ## Core Features & Use Cases - Persistent Three-File Planning: Creates and maintains task_plan.md, findings.md, and progress.md in your project directory as external working memory. - Automatic Context Injection: Hooks re-inject the current plan at each turn and before compaction, keeping goals in the model's attention window. - Parallel Plan Isolation: Supports multiple concurrent plans under .planning/<date>-<slug>/ directories with an active-plan pointer for multi-task workflows. - Plan Integrity Attestation: SHA-256 attestation locks an approved plan and blocks injection if the file is tampered with. - Use Case: Ask the agent to refactor a large module across many files; it writes a phased plan first, logs errors and decisions as it works, and resumes correctly after a /clear. ## Quick Start Ask the agent to plan out and execute your multi-step task using persistent planning files, for example: create a plan and implement the new authentication feature step by step.
