yt-title-optimizer
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data that could contain malicious instructions designed to influence agent behavior.
- Ingestion points: The skill gathers input from user-provided topics, draft titles, and external video URLs which are processed via
lib/url_parser.py(SKILL.md, Step 1). - Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the processed data.
- Capability inventory: The skill possesses read capabilities via
lib/url_parser.pyand write/publish capabilities vialib.publish(SKILL.md, Step 8). - Sanitization: While the skill includes a stylistic 'Humanizer pass' (Step 5), it lacks security-focused sanitization or validation of the input data to prevent instruction injection.
- [DATA_EXFILTRATION]: The skill performs network operations to fetch and process content from user-provided URLs.
- Evidence: Step 1 in
SKILL.mdspecifies the use oflib/url_parser.pyto parse arbitrary external URLs provided by the user. While necessary for the skill's function, this facilitates network access to non-whitelisted domains.
Audit Metadata