x-reply-drafter
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (tweet text and thread context) to generate drafts, creating a potential vector for malicious instructions embedded in that content to influence the agent's behavior.
- Ingestion points: The skill instructions in
SKILL.md(Step 2) explicitly prompt the user to paste external tweet content and thread history into the agent's context. - Boundary markers: There are no explicit delimiters or "ignore embedded instructions" warnings described for the interpolation of this untrusted content into the drafting prompts.
- Capability inventory: The skill uses
lib.publish(Step 7) to finalize the output, though the documentation notes this primarily generates a copy-paste block for the user to post manually. - Sanitization: No security-oriented sanitization or input validation is mentioned; the "humanizer pass" described is limited to stylistic and vocabulary adjustments.
Audit Metadata