code-search
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
Bashtool to run search commands. The execution templateprbot code search -q "<SEARCH_QUERY>"interpolates user-controlled strings directly into a shell command, which creates a potential surface for command injection if the input contains shell metacharacters.- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes code snippets and repository data from an external search service. \n - Ingestion points: Data enters the agent's context from the
comfy-codesearchservice output via theprbotcommand (SKILL.md). \n - Boundary markers: The skill does not define delimiters or instructions to ignore instructions embedded within search results. \n
- Capability inventory: The agent has access to the
Bashtool (SKILL.md), allowing for system-level execution. \n - Sanitization: There is no description of sanitizing or escaping the service output before processing.
Audit Metadata