review-changes
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external code changes and function definitions through tools like
detect_changes_toolandquery_graph_tool. This involves ingesting untrusted data which could potentially contain instructions intended to influence the agent's review logic or output. - Ingestion points:
detect_changes_toolandquery_graph_tooldefined inSKILL.mdprocess user-provided code changes. - Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded prompts within the code being reviewed.
- Capability inventory: The skill is limited to graph querying and analysis tools (
detect_changes_tool,get_affected_flows_tool,query_graph_tool,get_impact_radius_tool); it does not demonstrate capabilities for network exfiltration or direct file system modification. - Sanitization: There is no explicit mention of sanitizing or escaping the content retrieved from the code analysis tools before it is processed by the agent.
- [NO_CODE]: The skill consists exclusively of instructional markdown and YAML configuration. No executable scripts (Python, JavaScript, Shell) are included in the skill definition.
Audit Metadata