query-and-explain
Pass
Audited by Gen Agent Trust Hub on Jul 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes the 'graphify' CLI and standard shell utilities such as 'find' and 'grep' to query local knowledge graphs and perform file staleness checks.
- [PROMPT_INJECTION]: The skill includes 'Self-Evolving' and 'Post-Execution Reflection' instructions that direct the agent to autonomously modify its own instruction file (SKILL.md) and project documentation (CLAUDE.md) based on query results.
- [PROMPT_INJECTION]: An indirect prompt injection surface is present. Ingestion points: graphify-out/graph.json. Boundary markers: Absent. Capability inventory: Bash tool access and file-write permissions for the skill's own source files. Sanitization: No sanitization or validation of the graph data is specified before it is used to inform self-updates to the skill's instructions.
Audit Metadata