skills.rest
wolfstar-project
wolfstar-projectOFFICIAL
Explore all skills by this author →
WebsiteGitHub

planning-with-files

Maintains persistent markdown planning files on disk so agent work survives context loss and session resets.

Updated Aug 8, 2026
Search Tags:#planning#markdown#context management#hooks#ai agent#task tracking#session recovery
One-click install
npx skills add https://github.com/wolfstar-project/code-zero --skill planning-with-files-wolfstar-project
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: planning-with-files
Source: https://github.com/wolfstar-project/code-zero/tree/main/.agents/skills/planning-with-files
Command: npx skills add https://github.com/wolfstar-project/code-zero --skill planning-with-files-wolfstar-project

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? AI coding agents lose track of goals during long multi-step tasks because context windows are volatile and get compacted or cleared. This Skill stores task plans, findings, and progress as markdown files on disk, so work survives /clear, compaction, and session restarts. ## Core Features & Use Cases - Persistent three-file planning: Creates and maintains task_plan.md, findings.md, and progress.md in the project directory as external working memory. - Hook-based context injection: UserPromptSubmit, PreToolUse, PostToolUse, Stop, and PreCompact hooks re-inject plan state automatically, with optional SHA-256 attestation to block tampered plans. - Parallel and gated modes: Supports isolated per-task plan directories under .planning/, plus autonomous and gated modes with a completion gate, ledger summaries, and runaway guards. - Use Case: An agent running a 50-step refactor creates task_plan.md, logs each phase and error, and after a /clear resumes exactly where it left off by re-reading the plan files. ## Quick Start Ask the agent to plan out your multi-step project using planning files so it creates task_plan.md, findings.md, and progress.md before starting work.

Frequently Asked Questions about planning-with-files

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I keep an AI coding agent on track during long tasks?▼

Use file-based planning: the agent writes task_plan.md with phases, findings.md with discoveries, and progress.md with a session log. Hooks re-inject the plan into context each turn so goals stay in the attention window even after many tool calls.

How do I recover agent context after /clear or compaction?▼

Run the session-catchup.py script, which reads the previous session store and reports unsynced context. Then review git diff and the existing planning files, update them, and continue the task from the recorded phase.

Can multiple planning sessions run in the same repository?▼

Yes. init-session.sh with a name argument creates an isolated plan under .planning/<date>-<slug>/. Switch between plans with set-active-plan.sh or pin a terminal with the PLAN_ID or PWF_PLAN_ROOT environment variables.

Does planning-with-files work on Windows?▼

Yes. The skill ships PowerShell equivalents of its shell scripts, including attest-plan.ps1 and check-complete.ps1, and the Stop hook detects Windows environments to invoke the PowerShell gate automatically.

How does the skill prevent prompt injection through plan files?▼

Plan content is wrapped in BEGIN/END delimiters and treated as data, and an optional SHA-256 attestation locks the approved plan. If the file changes without re-attestation, hooks refuse injection and emit a [PLAN TAMPERED] warning.

When should I not use file-based planning?▼

Skip it for simple questions, single-file edits, and quick lookups. The pattern is designed for multi-step tasks of three or more steps, research work, and anything spanning many tool calls where state must persist.

Related Skills

Explore More Software Engineering Skills→
tanveerriaztanveerriaz
1

prd-to-build-plan

Converts PRDs and feature requests into phased build plans with acceptance criteria.

Community
Basic
pkkingpkking

planning-with-files

Maintain task_plan.md, findings.md, and progress.md for multi-step task planning.

Community
Advanced
CygnusfearCygnusfear
5

check-plan

Audit implementation progress against a plan and generate per-file TODOs.

Community
Basic
guanyangguanyang
960

planning-with-files

Maintains persistent markdown planning files on disk for multi-step AI agent tasks.

Community
Advanced