ai-anti-patterns
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a legitimate writing-quality tool focused on identifying 'AI tells' or anti-patterns in machine-generated or AI-assisted text. It follows established style guides (e.g., Wikipedia's AI cleanup guidelines) and uses statistical analysis to distinguish human prose from common AI defaults.
- [SAFE]: All included scripts (
style_metrics.py,scored-tics-patterns.py, and variouswikipedia-*.pyscripts) are transparency-focused linter implementations using standard Python regex and stylometric feature extraction (sentence length, diction, etc.). They operate locally on the user's draft files. - [SAFE]: No network activity, credential access, or privilege escalation patterns were found. The skill does not download external code; its reference files are static documentation or local regex definitions.
- [SAFE]: The skill uses an adversarial approach internally to detect AI-generated content (e.g., hallucinated DOIs, prompt refusals like 'As an AI language model'), which is its primary functional purpose. It does not contain prompt injections aimed at the agent analyzing it.
Audit Metadata