Skip to content
SkillModel-led4,089 tokens in use

Neat Freak

End-of-session cleanup that aligns project docs, CLAUDE.md or AGENTS.md rule files and agent memory with what the code and running services do today.

22,429 stars on the repository MIT LicenseKKKKhazix/khazix-skills · neat-freak
/plugin marketplace add KKKKhazix/khazix-skills /plugin install khazix-skills@khazix-skills
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Neat Freak costs

259tokens idle
4,089tokens in use
11,634tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works. Scripts: scripts/audit-inventory.sh.

Touches

shelldetected

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Fair
  • Frontmatter8/15
  • Structure8/25
  • Examples6/15
  • Output0/15
  • Size13/15
  • Resources15/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent
  • Runs shell commands−5

    does this on the machine, and the frontmatter does not say so

    scripts/audit-inventory.sh, line 4

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity80
  • Maintenance84
  • Community60
  • Maturity28
  • Momentum40

Neat Freak is the editor for a project’s written knowledge. After a work session it lists every rule file, README, doc and memory file, searches them for outdated names and claims, and corrects them in place so the next session or teammate reads one right answer.

It ranks evidence from current code and running services above docs and memory, audits whether stated project rules are followed, and reports what changed. Deleting files needs your authorisation, so it lists candidates with reasons.

What You Get

  • Project docs and rule files corrected to match the code, commands and ports that exist
  • Agent memory updated, merged or pruned, with lasting lessons moved to a durable doc first
  • A minimal rule file created when a working project has none
  • A closing report listing changes, deletion candidates awaiting your approval, unresolved conflicts and unverified claims

Common Use Cases

  • Syncing documentation after several rounds of code changes
  • Cleaning conflicting or expired entries from CLAUDE.md or agent memory
  • Preparing a clean handoff for a new teammate
  • Auditing whether the rules written in a project are being followed

What Makes It Unique

It treats current code and live services as the authority and everything written as a lead to verify, and it demands deployment evidence before any file says a change is live. It edits like an editor and does not append dated notes. It also protects your work: machine-generated memory stays read-only, deletion needs your authorisation, and instructions found inside project files are never treated as authorisation. The skill ships evaluation fixtures, and its fake projects are not real repositories to clean.

Under The Hood

The agent starts by running the bundled inventory shell script, which only reads metadata and paths: agent home folders, rule files up the directory tree, git status and every Markdown file. It then reads rule files, docs and memory, searches the project for old names, and edits those files directly. It may run project commands, tests and link checks to verify claims, and it can create or delete files only within the authorisation you give. Subagents can handle read-only inventory work in parallel.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers