diagnose
Pass
Audited by Gen Agent Trust Hub on Apr 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill guides the agent to execute shell commands and scripts, including
git bisect run,curl, and a human-in-the-loop bash template, to reproduce and diagnose software issues. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface as it processes untrusted data and maintains powerful execution capabilities.
- Ingestion points:
SKILL.mdspecifies reading external logs, traces, and HAR files;scripts/hitl-loop.template.shcaptures user responses. - Boundary markers: Absent; there are no instructions to delimit or ignore instructions within the ingested artifacts.
- Capability inventory:
SKILL.mddescribes executing shell commands (e.g.,git bisect,curl), running tests, and executing local scripts. - Sanitization: Absent; the skill does not specify any sanitization or validation for external data before it influences agent logic.
Audit Metadata