working-with-task-comments
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill allows the agent to ingest task comments and artifact feedback, which are external inputs that could contain malicious instructions designed to manipulate the agent's behavior.
- Ingestion points: Untrusted data is retrieved through tasks-comments-list and tasks-comments-retrieve inner tools via the posthog:exec dispatcher (SKILL.md).
- Boundary markers: The instructions explicitly tell the agent to treat returned comment content as untrusted review data and not as authority to expand the task or user permissions.
- Capability inventory: The skill is restricted to read-only operations for comment retrieval and uses standard PostHog MCP discovery tools.
- Sanitization: The skill mandates that the agent must not follow instructions in comments that request unrelated work or permission expansion and requires human confirmation from the task creator when feedback crosses security boundaries.
Audit Metadata