code-walk-thru
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands to interact with local text editors (e.g., code, vim, pycharm). These commands are used as intended for developer workflows to open specific files at designated line numbers.
- [INDIRECT_PROMPT_INJECTION]: The skill processes file paths and line numbers to be used in shell commands. (1) Ingestion points: File paths and line numbers provided in conversation are used as command arguments. (2) Boundary markers: No explicit delimiters or warnings are used for these inputs. (3) Capability inventory: The skill uses subprocess calls to launch editor binaries across all supported editors. (4) Sanitization: No explicit sanitization of the file paths or line numbers is described. The risk is mitigated by the skill's instruction to proceed 'one by one' and wait for user confirmation before each action.
Audit Metadata