Skillget

sloptrim

agent
Security Audit
Fail
Health Pass
  • License — License: Apache-2.0
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 147 GitHub stars
Code Fail
  • fs module — File system access in hooks/sloptrim-activate.js
  • process.env — Environment variable access in hooks/sloptrim-guard.js
  • fs module — File system access in hooks/sloptrim-guard.js
  • child_process — Shell command execution capability in hooks/sloptrim-lib.js
  • fs.rmSync — Destructive file system operation in hooks/sloptrim-lib.js
  • os.homedir — User home directory access in hooks/sloptrim-lib.js
  • process.env — Environment variable access in hooks/sloptrim-lib.js
  • fs module — File system access in hooks/sloptrim-lib.js
  • fs module — File system access in hooks/sloptrim-stats.js
  • child_process — Shell command execution capability in hooks/sloptrim-tracker.js
  • fs module — File system access in hooks/sloptrim-tracker.js
Permissions Pass
  • Permissions — No dangerous permissions requested

No AI report is available for this listing yet.

SUMMARY

A local detector for AI-writing patterns. Scores every prose file your agent saves. Python standard library only, no network, no model.

README.md

sloptrim

A local detector for AI-writing patterns. It scores every prose file your agent
saves and asks for the flagged spans to be fixed.
Python standard library only, no network, no model. Prose only, never code.

test
Version
License
Dependencies
Tests
Python

Install · What it does · Measured · Limits · Patterns · Ethics

[!IMPORTANT]
This is a command-line tool and an agent plugin. There is no website and no hosted version.
Nothing you write is uploaded, there is no account, and no text ever leaves your machine.
Any site offering a service under this name is unrelated to this project.

A recorded session. The contract arrives at session start, again on the prompt, and again inside a subagent. A saved file scores 45 and its patterns are named. After the fix the next read is silent at 0. The same text saved as a .docx scores the same, read out of the zip. Every hook latency is measured.
A recorded session, not a mock-up. record/session_capture.py fires the real hooks with the payloads Claude Code sends and writes record/session.json; record/anim.py draws it. The hook lines, the scores and the milliseconds come from that capture; the grey labels beside them are narration. The prose being scored is record/draft.md, a short piece written for the recording.

Install

Paste into Claude Code, Codex, Cursor, or any coding agent:

Install the sloptrim plugin from https://github.com/seyedehsanhadi/sloptrim

Restart, then run /sloptrim doctor. It answers with four [OK] lines.

Explicit commands, and installing without the marketplace 
/plugin marketplace add seyedehsanhadi/sloptrim
/plugin install sloptrim@sloptrim

git clone https://github.com/seyedehsanhadi/sloptrim.git ~/.claude/skills/sloptrim
mkdir -p ~/.claude/commands
cp ~/.claude/skills/sloptrim/install/sloptrim-command.md ~/.claude/commands/sloptrim.md

Do not skip the mkdir. On a fresh machine ~/.claude/commands does not exist yet
and the copy fails with "No such file or directory". In PowerShell the last two lines
are New-Item -ItemType Directory -Force $HOME/.claude/commands and Copy-Item.
The copy puts /sloptrim in the / menu, because Claude Code does not scan a skill
folder's own commands/. A marketplace install needs no such step. Either way the
router also answers to /sloptrim:sloptrim.

What it does

The score is 0-100 against 71 documented patterns. 62 of them have a detector; the
other 9 need a reading and are worked during the rewrite. Of the 62, 50 can move the
score and 12 are reported as writing advice and count for nothing: most of them because
measurement showed they mark formal register rather than machine authorship, the rest
because they are typographic habits.
Formats 20, including .docx, .pptx, .xlsx, OpenDocument, .epub, .ipynb, LaTeX
Runs in Claude Code, on save. Other agents via /sloptrim init, which writes the contract to AGENTS.md, and .cursor/rules/
Needs Node for the hooks, Python 3.9 or newer for the detector, nothing else
Suite 101 Python tests and 57 hook checks, green in CI on Linux, Windows and macOS, against Python 3.9 and 3.13
Does not see A file written by a Bash command, which reaches disk without passing Write or Edit
Command Effect
/sloptrim full Contract + guard (default)
/sloptrim strict Flags at 20 instead of 40, and asks for a character scrub
/sloptrim lite / off Contract only / nothing
/sloptrim check <file> Score a file, name the tells, no rewrite
/sloptrim init Write the contract to ./AGENTS.md and a Cursor rule to ./.cursor/rules/
/sloptrim doctor Diagnose the install 
python scripts/detect.py draft.docx    # JSON: patterns, metrics, 0-100 score

A score lands in one of five bands: clean, light tells, mixed, heavy tells,
pervasive tells. The guard nudges above 40, or above 20 in strict mode.

Measured

The figures below were measured against corpora held privately. Neither those
corpora nor the harness that read them is in this repository, and nothing here
recomputes any of it. They are cited as results, with the corpus named, and cannot
be re-derived from what you have cloned.

False positives on human prose, at the default threshold, worst corpus first:

corpus n rate
American textbooks, 30 titles 9,333 0.85%
MAGE human web text 504 0.40%
PubMed abstracts, pre-2020 529 0.00%
arXiv abstracts, pre-2021 939 0.00%

Detection, ROC-AUC per corpus. The worst one is the headline:

Detection by corpus, ROC-AUC: American textbooks 0.963, RAID 0.791, MAGE 0.669, a frontier model 0.551. 0.5 is a coin flip.

What it cannot do

It cannot tell you whether a current frontier model wrote something. The machine
arms above come from GPT-2, GPT-J, OPT, FLAN-T5, MPT, Mistral, Mixtral, GPT-3.5 and
GPT-4, nothing newer. A run against a current frontier model measured ROC-AUC 0.551,
close to a coin flip.

The writing contract has no measured effect. Twenty documents drafted twice from
one brief with the switch toggled, both arms verified from the transcripts: mean
change +2.25, bootstrap 95% CI -5.65 to +9.90, sign test p = 0.27. Its banned-word
list works; nothing else in it does. That is 20 pairs against the 40 the protocol asks
for, so the question is unresolved rather than settled.

It is not an authorship classifier and must not be used as one. A score says
something about writing, never about a person. Read ETHICS.md.

License

Apache-2.0 (full text, NOTICE). Cite with
CITATION.cff.

Reviews (0)

No results found