kimi-skills-finder
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use shell commands (specifically
ls) to list files in/app/.agents/skills/and/app/.user/skills/. This is used to verify the existence of local skills and is consistent with the skill's stated purpose. - [INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes information from external, untrusted sources including GitHub repositories and the SkillHub website.
- Ingestion points: Data is ingested through
web_searchresults (GitHub READMEs/metadata) andbrowser_visit(SkillHub page content). - Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within the fetched skill descriptions.
- Capability inventory: The agent has access to
shellexecution,web_search, andbrowser_visittools. - Sanitization: There is no mention of sanitizing, escaping, or validating the content retrieved from external sources before the agent evaluates it.
Audit Metadata