Simplify Codebase Safely
Maps every dependency of a file, skill, or codebase, then removes only cuts that tests prove safe.
- Repository
- danielmiessler/LifeOS
- Stars
- 19,682
- Licence
- MIT
- Files
- 4 files
What it does
Once installed, the agent handles requests to make a file, tool, skill, subsystem, or repo smaller and clearer. It first surveys the target read-only, finding everything that calls or references it (imports, hooks, settings, docs, tests). It then classifies each element as load-bearing, duplicate, dead, cruft, or over-abstracted, proposes ranked cuts for your approval, and only after approval applies them, checking before-and-after parity so nothing a dependent relies on breaks.
Great for
- Developers with a bloated repo or tool who want cleanup without breakage risk
- Maintainers who need evidence of what is actually unused before deleting
- Teams that want an audit trail of each cut recorded in work folders
Skip it if
- Always-on context files such as CLAUDE.md, which the skill routes to a different skill (Trim) (wrong tool for the job)
- Cleaning up only the current diff or finding bugs, which other tools handle better (scope mismatch)
- Projects without tests or probes, since parity cannot be proven and the skill's safety claim weakens (limited verification)
Try asking
- “simplify LIFEOS/TOOLS/ReferenceCheck.ts”
- “what can we delete in this subsystem?”
- “this has gotten bloated, map it before we cut anything”
- “remove duplication in this skill folder”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 4 original files, the MIT licence and an attribution note.
Simplify.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip Simplify.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip Simplify.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Provide the target's tests or baseline checks up front, since parity proof depends on them.
- Approve proposed cuts by row number (for example, y but skip 4) rather than all at once.
- Expect zero search hits to be insufficient: dynamic and string-based references need the skill's reference-check approach.
What's inside
- Intents/Apply.md
- Intents/Propose.md
- Intents/Survey.md
- SKILL.md
Redistributed unmodified from danielmiessler/LifeOS under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.