fb-repurposer
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content (LinkedIn posts, X threads, blogs) and interacts with a publishing tool, which introduces a risk of the agent following malicious instructions embedded in the source data.\n
- Ingestion points: Ingests user-provided source content in various formats including text posts, transcripts, and URLs (Steps 1 and 2 in SKILL.md).\n
- Boundary markers: Absent. There are no explicit instructions to use delimiters or specific warnings to ignore embedded instructions within the provided source content.\n
- Capability inventory: The skill utilizes the
lib.publishtool to post content to Facebook (Step 10 in SKILL.md).\n - Sanitization: The skill implements an 'Approval card' (Step 9) that requires explicit user confirmation before any content is published, providing a robust human-in-the-loop checkpoint.
Audit Metadata