posthog
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill acknowledges an inherent attack surface where data retrieved from the PostHog platform (such as application logs, session recording summaries, and LLM trace content) could contain adversarial instructions. The skill explicitly mitigates this risk by providing clear instructions for the agent to treat such data as evidence rather than actionable commands.
- Capability inventory: The skill allows the agent to perform high-impact actions including modifying feature flags, launching experiments, and updating data warehouse models, which heightens the risk if injection occurs.
- Ingestion points: Data enters the agent's reasoning loop via tools such as
query-logs,query-llm-trace,surveys-responses-list, andvision-observations-list. - Sanitization: The skill mandates that the agent verify all findings across multiple independent signals and prohibits auto-execution of actions without human confirmation.
Audit Metadata