diagnose-issue
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes external, untrusted content from codebase files, bug reports, and logs which could contain malicious instructions designed to influence the agent's analysis.
- Ingestion points: User-provided issue statements, bug reports, Jira tickets, codebase files, and log files as specified in the Workflow section of
SKILL.md. - Boundary markers: Absent. The instructions do not provide explicit delimiters or warnings to the agent to disregard instructions potentially embedded within the data it analyzes.
- Capability inventory: The skill has the capability to search the codebase, read arbitrary files, and execute local runtime commands as described in the Code Map section of
SKILL.md. - Sanitization: Absent. There are no instructions for validating, escaping, or filtering the content found in external data sources.
- [COMMAND_EXECUTION]: The skill explicitly instructs the agent to use runtime commands for diagnosis.
- Evidence:
SKILL.mdcontains the instruction "Use runtime commands only when they are cheap, local, and useful for diagnosis." While intended for debugging (e.g., running tests or listing files), this capability could be exploited if an attacker embeds commands in a file the agent is instructed to analyze.
Audit Metadata