spec-kitty-implement
Pass
Audited by Gen Agent Trust Hub on Apr 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the spec-kitty CLI tool and git to handle workspace creation and task management. These operations are consistent with the documented software development workflow.\n- [PROMPT_INJECTION]: The workflow involves an indirect prompt injection surface where the agent processes large-scale instructions (~1000 lines) emitted by the spec-kitty tool output.\n
- Ingestion points: Standard output of the
spec-kitty agent workflow implementcommand.\n - Boundary markers: Not specified; the agent is instructed to read the entire output to find completion commands.\n
- Capability inventory: The agent is authorized to modify files within worktrees and execute version control commands.\n
- Sanitization: The instructions do not describe any sanitization or verification of the external tool output before the agent acts upon it.
Audit Metadata