fiction-humanizer-zh
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of editorial instructions, a local Python analysis script, and reference documentation for stylistic refinement. No malicious behavior, obfuscation, or data exfiltration was found.
- [COMMAND_EXECUTION]: The skill includes a Python script
scripts/check_fiction_prose.pywhich performs local text analysis using regular expressions and basic statistics. It uses standard libraries (argparse, json, re, statistics, pathlib) and does not perform network operations or access sensitive system files. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided fiction text. While this is an ingestion point for untrusted data, the skill acts as a text refiner without dangerous tool capabilities, posing no significant injection risk.
- [EXTERNAL_DOWNLOADS]: The documentation references external projects for attribution purposes (e.g., xiaoyuCR7/long-novel-skill, Ruixu-dev/human-flavor-engine). These are provided as plaintext citations and are not utilized for remote code execution or automated downloads.
Audit Metadata