slo-investigate
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection vulnerability surface by processing external, potentially untrusted content.
- Ingestion points: In
SKILL.md(Step 6), the skill extracts runbook URLs from SLO definition annotations and retrieves external markdown files from GitHub repositories using thegh apicommand. - Boundary markers: No delimiter boundaries or isolation instructions are provided when rendering or analyzing the retrieved external runbook content.
- Capability inventory: The skill utilizes the
Bashtool to run administrative commands likegcx metrics query,gcx alert rules list, andgcx slo definitions status. - Sanitization: There is no input sanitization or filtering applied to the fetched runbook content before it is included in the output report provided to the agent and user.
Audit Metadata