The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified. The skill correctly utilizes first-party tools within the 'posthog' namespace to perform legitimate analytics and debugging tasks.
[DATA_EXPOSURE]: The skill is designed to retrieve and process sensitive session information, including person properties (email, name) and browser events. This access is necessary for the skill's primary function and occurs entirely within the authenticated PostHog environment.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from session events and AI-generated summaries. While this represents an attack surface, the risk is mitigated by the structured nature of the investigation workflow and the use of predefined SQL templates. This behavior is standard for session analysis tools.
Ingestion points: posthog:execute-sql (event properties) and posthog:vision-observations-get (AI summary content) in SKILL.md.
Boundary markers: Absent.
Capability inventory: Read-only access to session, person, and event data; SQL query execution.
Sanitization: None identified; relies on the underlying platform's handling of session data.

investigating-replay