kimi-result-handling
Pass
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill establishes strictly defensive presentation rules for handling external model output.
- [SAFE]: Implements a "Never auto-execute" policy (Rule 3 in SKILL.md) which explicitly prevents the agent from running commands or applying code changes suggested in free-text output without user confirmation.
- [SAFE]: Instructs the agent not to parse free-text output from the Kimi provider as executable shell commands, providing a significant safeguard against indirect prompt injection.
- [SAFE]: Mentions that tool calls (like
apply_patch) are handled via a distinct channel with its own authorization model, ensuring that privileged actions remain under user control.
Audit Metadata