draft-tweet
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes the 'xurl' CLI to check user identity and fetch recent posts, and uses 'no-tropes' to refine draft options.- [DATA_EXFILTRATION]: The skill retrieves the user's recent X/Twitter posts for linguistic style analysis. This data access is used locally for analysis and does not involve unauthorized external transmission.- [PROMPT_INJECTION]: The skill processes user-generated tweet history as input (Ingestion points: xurl command output; Boundary markers: Absent; Capability inventory: xurl CLI; Sanitization: Absent). This creates a surface for indirect prompt injection, but the instructions to match style rather than content mitigate potential exploitation.
Audit Metadata