kata-audit-milestone
Pass
Audited by Gen Agent Trust Hub on Apr 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes various shell commands such as find, grep, cat, and node to inspect project files and directories. It also invokes a local vendor-provided script (scripts/kata-lib.cjs) for configuration and migration checks.\n- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface by ingesting data from *-SUMMARY.md and *-VERIFICATION.md files (Ingestion points: SKILL.md steps 2 and 3) and interpolating it into the instructions for a subagent (Task prompt in SKILL.md step 3). There is an absence of boundary markers or delimiters (Boundary markers: absent) around the interpolated content and no sanitization (Sanitization: absent) of the file contents before they are passed to the subagent. The subagent has capabilities including codebase inspection via shell commands (Capability inventory: defined in integration-checker-instructions.md).
Audit Metadata