planning-with-files_skill

This skill helps you plan, break down, and organize complex tasks using on-disk planning files and templates to improve traceability.
  • Python

335

GitHub Stars

3

Bundled Files

2 months ago

Catalog Refreshed

3 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill guanyang/antigravity-skills --skill planning-with-files

  • examples.md4.3 KB
  • reference.md7.9 KB
  • SKILL.md6.9 KB

Overview

This skill implements Manus-style file-based planning to organize and track progress on multi-step projects. It creates and maintains three canonical files — task_plan.md, findings.md, and progress.md — in your project directory to persist decisions, discoveries, and session logs. Use it to structure work that requires many tool calls, repeatable recovery, and clear error tracking.

How this skill works

Before doing substantive work the skill checks for unsynced context from previous sessions and advises on catch-up steps. It generates templates for task_plan.md, findings.md, and progress.md in the project root and enforces a workflow: plan first, write discoveries frequently, update progress continuously, and log all errors. The skill also supports automatic session recovery so agents can resume safely after a cleared or interrupted session.

When to use it

  • Planning multi-step projects (3+ phases)
  • Research or discovery workflows that produce many findings
  • Tasks that require more than five tool calls or external interactions
  • Long-running engineering or design efforts needing persistent state
  • Any work where reproducible session recovery and error tracking matter

Best practices

  • Always create task_plan.md before executing steps and re-read it before major decisions
  • Follow the 2-action rule: after two view/search actions, immediately save key findings to files
  • Log every error and update the plan with status changes after each phase
  • Keep planning files in the project root, not inside the skill installation folder
  • After encountering repeated failures, follow the 3-strike protocol and escalate to the user if needed

Example use cases

  • Break down and track a multi-week feature development with phases, milestones, and decisions recorded
  • Run complex research that accumulates notes and references in findings.md for reproducibility
  • Manage a data pipeline migration with progress.md recording tests, results, and rollbacks
  • Coordinate a design iteration where decisions and discovered constraints are persisted for later review
  • Resume work after an interruption by running the session-catchup script and reconciling git diffs with planning files

FAQ

It creates task_plan.md, findings.md, and progress.md in your project directory (not the skill folder). Templates live with the skill but files belong in the project root.

How does session recovery work?

Run the provided session-catchup script to detect unsynced context, then read planning files and reconcile changes (git diff recommended) before continuing.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational