preflight-new-skill
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DATA_EXPOSURE]: Accesses a specific configuration file located at
~/.config/preflight-new-skill/leak-patterns.txtto load custom search patterns for auditing purposes. - [DATA_EXPOSURE]: Scans files within the repository for sensitive patterns, including internal file paths, private hostnames, and regex-matched credential formats (e.g., AWS, GitHub, and Slack tokens), to identify and prevent accidental public disclosure.
- [INDIRECT_PROMPT_INJECTION]: Processes user-controlled content in
SKILL.mdfiles as part of the audit workflow. - Ingestion points: Reads newly added or modified
SKILL.mdfiles identified via git. - Boundary markers: None identified.
- Capability inventory: Uses shell-based utilities like
grepandwcto perform analysis on the ingested content. - Sanitization: None; performs pattern matching directly on file content.
- [COMMAND_EXECUTION]: Utilizes shell commands including
git diff,grep,wc,sort, andtrto automate the auditing of repository changes and file structure.
Audit Metadata