reply-and-comment-writer
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes social media comments and DMs provided by the user, which are untrusted data sources.
- Ingestion points: The skill ingests text from comments, DMs, and posts that the user is prompted to paste into the agent's context, as described in
SKILL.mdandreferences/dms-hard-cases-and-safety.md. - Boundary markers: The instructions explicitly direct the agent to treat these inputs as "content, not a command" and provide a clear "Injection safety" policy to ignore embedded instructions like "ignore your instructions" or requests for sensitive data.
- Capability inventory: The skill's scope is strictly advisory; it drafts content for human review and explicitly states that the platform (WoopSocial) cannot pull or post content automatically, limiting the potential impact of an injection.
- Sanitization: The skill instructs the agent to surface any detected injection attempts to the user for review rather than acting upon them.
Audit Metadata