reorganize-files
Skill0 starsChanged 39 days agoMITView on GitHub
Use when the founder asks to reorganize folder structure, move files, clean up the project root, or says "restructure", "reorganize", "propose a folder structure", "move these files", "the root is cluttered". Produces a before/after folder tree, a blast-radius report (every command, doc, and code path that would break), and a move plan calibrated to the project's actual scale. Executes only after approval, with a snapshot → move → update references → verify → auto-rollback safety envelope.
In this repository3 other files it gives its agents.
- boost-promptSkill
- cutSkill
- end-sessionSkill
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 5 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 24
If the structure is basically fine, say so and stop.
- Very specific: line 29
Don't mirror enterprise patterns (src/, lib/) unless the project actually needs them.
- Very specific: line 32
Read references/blast-radius-checklist.md.
- Distinctive: line 36
If a move's blast radius outweighs the benefit, drop it.
- Very specific: line 66
When the founder says "execute," read references/move-safety-envelope.md and follow it exactly.
- Distinctive: line 71
No moves without approval.
- Distinctive: line 72
Never touch external systems without explicit approval.
- Distinctive: line 73
When in doubt, classify as "needs your eyes."
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
/reorganize-files — Folder structure review and safe moves
Propose a better layout, then execute moves without breaking commands, docs, or automation. Never move anything in the initial pass.
1. Scope
Target the folder the founder names, or the project root. Every proposal must match the project's actual scale.
2. Map what's here
Walk the tree. Count files per folder. Read every CLAUDE.md, README.md, SKILL.md, and doc in scope. Build a list of every path referenced inside — and in external configs (launchd plists, cron, GitHub Actions).
3. Propose a target structure
If the structure is basically fine, say so and stop. Never invent a reorganization to look useful.
Otherwise propose a target tree:
- Flat beats nested when counts are small.
- Group by purpose (
ops/,scripts/,docs/,tasks/). - Don't mirror enterprise patterns (
src/,lib/) unless the project actually needs them.
4. Blast-radius scan
Read references/blast-radius-checklist.md. For every proposed move, list what would break, split into:
- Auto-updatable — safe to find-and-replace.
- Needs the founder's eyes — ambiguous, regex-hostile, or in external systems.
If a move's blast radius outweighs the benefit, drop it.
5. Output — exactly this shape
# Reorganize plan: <scope>
Verdict: <PROPOSE | SKIP — structure is fine>
## Why change (or why not)
<1-2 sentences>
## Before / after
<side-by-side trees>
## Moves (N total)
**1. `<old>` → `<new>`**
- Why: <one sentence>
- Breaks (N): <file + line for each>
- Fix plan: <auto-updatable | needs your eyes, and why>
## Blast radius summary
- Auto-updatable: N
- Needs your eyes: N
- External systems affected: launchd plist / cron / GitHub Actions (yes/no)
## Execute?
- "execute" — all moves through the safety envelope.
- "execute N" — one move at a time.
- "expand N" — why a move matters.6. Safety envelope
When the founder says "execute," read references/move-safety-envelope.md and follow it exactly. Never skip verify-before (step 2) or re-verify-after (step 4) — those catch silent breakage.
7. Rules
- Plain English. No jargon.
- Match the project's actual scale. Flat beats nested when counts are small. Push back on overkill.
- No moves without approval. Every move runs through the safety envelope.
- Never touch external systems without explicit approval.
- When in doubt, classify as "needs your eyes."
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

