query-examples
Pass
Audited by Gen Agent Trust Hub on Apr 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of reference material and SQL query examples for PostHog data. No malicious patterns such as prompt injection, unauthorized network calls, or credential exfiltration were detected. External references trace back to the vendor's own infrastructure (posthog.com).
- [COMMAND_EXECUTION]: The guidelines suggest using bash commands to process large output files generated from analytical queries. This is presented as a strategy to handle data volume efficiently within the agent's environment while avoiding flooding the conversation context.
- [PROMPT_INJECTION]: The skill processes untrusted analytics data from external sources, presenting a surface for indirect prompt injection.
- Ingestion points: Analytics data enters the agent context via the posthog:execute-sql tool when querying the events or persons tables.
- Boundary markers: The guidelines instruct the agent to use explicit time range filters and schema verification steps before executing broad queries.
- Capability inventory: The skill utilizes the posthog:execute-sql tool and references the availability of shell tools for file exploration.
- Sanitization: The documentation includes strict rules against selecting the full properties JSON object and recommends cherry-picking specific fields to minimize exposure of unvalidated data.
Audit Metadata