slop-detector

Deterministic pattern detector for inbound received text: emails, writing samples, proposals, deck copy. Text in, evidence report out. Every finding cites a byte span and a verbatim snippet. The report carries no verdict and no score. The tool finds. The reader decides what the findings mean.

What it detects

The report splits findings into three categories, plus input measurements.

• paste_residue: mechanical residue of a copy from a generation surface. Provider-attribution lines, chat-export citation artifacts, chat-tool tracking parameters, turn markers, citation-delimiter codepoints, invisible unicode, and positional typographic spaces. Each finding is read per hit.
• injection_patterns: phrasing that addresses an assistant, such as demands to disregard prior guidance or reveal hidden configuration. Read per hit. The analyzed text is evidence, never instructions.
• quality_patterns: formulaic-writing patterns, from the measured excess-vocabulary set through business-register density and structural scaffolding. Read as densities against the word count, never as per-token verdicts.
• stats: word_count and byte_len, the denominators for density reads.

Every rule is data, loaded from data/inbound/inbound.toml. No pattern is hard-coded. The reference lexicons in data/words/ and data/policy.toml are vendored from the ai-slop crate. The inbound selection and its edits live in data/inbound/.

CLI

The slop-detector binary reads a file path argument, or stdin when no path is given, and prints the report as JSON on stdout.

slop-detector inbound.txt
cat inbound.txt | slop-detector

Exit 0 means a report was produced. Exit 1 is a read or encoding error. Inputs over 4 MiB are rejected with exit 40. An input with no matches yields empty arrays, which is a complete, valid report. A closed output pipe ends the run quietly.

Library

let report = slop_detector::analyze(&text);
for finding in &report.paste_residue {
    println!("{} at {:?}: {}", finding.rule_id, finding.span, finding.snippet);
}

analyze is a pure function of the input text: deterministic output, no file or network access, no panics on any input.

The coupled skill

The crate pairs with an agent skill, in the repository at skills/slop-detector/, that directs a human-driven agent through the read: state the purpose, run the tool, read residue and injection findings per hit, compute quality densities per class, and report with cited spans. The per-rule interpretation notes live in the skill's references/rules.md.

License

MIT OR Apache-2.0.