investigating-error-issue
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions direct the agent to ingest untrusted data from the user and interpolate it into high-capability tool calls.
- Ingestion points: User-provided issue IDs, URLs, and session IDs are used to drive the investigation process in SKILL.md.
- Capability inventory: The skill makes extensive use of the
posthog:execute-sqltool to query event and log data, as well asposthog:query-logsfor server-side analysis. - Boundary markers: There are no explicit instructions or delimiters defined to separate user input from the rest of the query logic, nor are there warnings to the agent to validate the structure of the input.
- Sanitization: The workflow demonstrates direct string substitution for values such as
<issue_id>and<session_id_from_step_2>inside SQL WHERE clauses (e.g.,WHERE issue_id = '<issue_id>'), which is a common pattern for injection surfaces.
Audit Metadata