kata-delegate
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill uses clear instructional language to define boundaries and does not contain patterns intended to bypass safety filters or override system instructions.
- [DATA_EXFILTRATION]: No commands accessing credentials, sensitive files, or making unauthorized network requests were identified.
- [PRIVILEGE_ESCALATION]: The skill explicitly disclaims authority for sensitive actions like merging, publishing, or sending messages, requiring explicit user or host permission for such actions.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a protocol for handling untrusted work requests (SKILL.md). It uses boundary markers like explicit exclusions and resource limits (SKILL.md). Its capabilities include starting workers via host monitoring tools (SKILL.md). Sanitization is implemented by preserving supplied exclusions in the output (SKILL.md).
- [COMMAND_EXECUTION]: No unauthorized shell commands or system-level operations are present in the skill instructions or reference files.
Audit Metadata