Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Uses persistent markdown files for general planning, progress tracking, and knowledge storage (Manus-style workflow). Use for multi-step tasks, research projects, or general organization WITHOUT mentioning PRD. For PRD-specific work, use prd-planner skill instead.
.claude/skills/planning-with-files/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-14 | ✗→✓ | ▲ Improved | 29% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 14% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -12% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 10% | 0% |
> "Work like Manus" — Uses persistent markdown files for planning, progress tracking, and knowledge storage.
A skill that transforms your workflow to use persistent markdown files for planning and progress tracking.
Claude Code (and most AI agents) suffer from:
For every complex task, create THREE files:
texttask_plan.md → Track phases and progress notes.md → Store research and findings [deliverable].md → Final output
text1. Create task_plan.md with goal and phases 2. Research → save to notes.md → update task_plan.md 3. Read notes.md → create deliverable → update task_plan.md 4. Deliver final output
Use this pattern for:
Skip for:
This skill is typically installed globally at ~/.claude/skills/planning-with-files/.
From this repository:
bashln -s /path/to/agent-playbook/skills/planning-with-files ~/.claude/skills/planning-with-files
If you prefer the standalone workflow, see the upstream repository in the Links section.
| Principle | Implementation | |-----------|----------------| | Filesystem as memory | Store in files, not context | | Attention manipulation | Re-read plan before decisions | | Error persistence | Log failures in plan file | | Goal tracking | Checkboxes show progress | | Append-only context | Never modify history |
You: "Research the benefits of TypeScript and write a summary"
Claude creates:
markdown# Task Plan: TypeScript Benefits Research ## Goal Create a research summary on TypeScript benefits. ## Phases - [x] Phase 1: Create plan ✓ - [ ] Phase 2: Research and gather sources (CURRENT) - [ ] Phase 3: Synthesize findings - [ ] Phase 4: Deliver summary ## Status **Currently in Phase 2** - Searching for sources
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-14 | fail→pass | 13,635 | 14,410 | +6% | 1 | 1 | 0% | 2,466 | 3,178 | +29% | 0 | 0 | — |
case-01 | fail→pass | 35,736 | 36,465 | +2% | 1 | 1 | 0% | 6,013 | 6,828 | +14% | 0 | 0 | — |
case-02 | fail→fail | 28,218 | 5,069 | -82% | 1 | 1 | 0% | 6,186 | 885 | -86% | 0 | 0 | — |
case-03 | fail→fail | 33,233 | 4,416 | -87% | 1 | 1 | 0% | 6,171 | 838 | -86% | 0 | 0 | — |
case-04 | pass→pass | 2,027 | 2,014 | -1% | 1 | 1 | 0% | 269 | 943 | +251% | 0 | 0 | — |
case-05 | fail→fail | 4,528 | 3,350 | -26% | 1 | 1 | 0% | 842 | 1,007 | +20% | 0 | 0 | — |
case-06 | pass→pass | 8,896 | 6,120 | -31% | 1 | 1 | 0% | 1,793 | 1,768 | -1% | 0 | 0 | — |
case-07 | fail→pass | 14,792 | 9,225 | -38% | 1 | 1 | 0% | 2,621 | 2,311 | -12% | 0 | 0 | — |
case-08 | fail→fail | 18,271 | 12,304 | -33% | 1 | 1 | 0% | 3,281 | 1,923 | -41% | 0 | 0 | — |
case-09 | fail→fail | 14,914 | 9,231 | -38% | 1 | 1 | 0% | 2,501 | 1,807 | -28% | 0 | 0 | — |
case-10 | fail→pass | 16,115 | 11,947 | -26% | 1 | 1 | 0% | 3,352 | 2,811 | -16% | 0 | 0 | — |
case-11 | fail→pass | 15,259 | 13,579 | -11% | 1 | 1 | 0% | 2,863 | 3,143 | +10% | 0 | 0 | — |
case-12 | fail→pass | 4,414 | 3,801 | -14% | 1 | 1 | 0% | 686 | 1,230 | +79% | 0 | 0 | — |
case-13 | fail→fail | 10,151 | 34,349 | +238% | 1 | 1 | 0% | 1,645 | 2,681 | +63% | 0 | 0 | — |
case-15 | fail→fail | 6,322 | 4,078 | -35% | 1 | 1 | 0% | 1,042 | 861 | -17% | 0 | 0 | — |
case-16 | fail→fail | 24,328 | 9,673 | -60% | 1 | 1 | 0% | 3,996 | 1,706 | -57% | 0 | 0 | — |
case-17 | pass→pass | 10,403 | 5,223 | -50% | 1 | 1 | 0% | 1,620 | 1,290 | -20% | 0 | 0 | — |
case-18 | pass→pass | 12,578 | 7,968 | -37% | 1 | 1 | 0% | 2,090 | 2,104 | +1% | 0 | 0 | — |
case-19 | fail→fail | 14,465 | 5,685 | -61% | 1 | 1 | 0% | 1,360 | 1,549 | +14% | 0 | 0 | — |
case-20 | fail→pass | 10,376 | 3,378 | -67% | 1 | 1 | 0% | 1,842 | 1,189 | -35% | 0 | 0 | — |
case-21 | fail→pass | 11,189 | 14,712 | +31% | 1 | 1 | 0% | 1,981 | 2,007 | +1% | 0 | 0 | — |
case-22 | pass→pass | 1,771 | 1,528 | -14% | 1 | 1 | 0% | 201 | 787 | +292% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 20 counted toward the lift figure. The other 2 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +36 percentage points is the difference between those two pass rates over the 20 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 7/24/2026 | — |
Other measured skills in the registry, with their headline benchmark lift.