clawhub
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches the
clawhubCLI utility from the npm registry and downloads additional skill content from the vendor's registry athttps://clawhub.com.\n- [COMMAND_EXECUTION]: Executes system commands for thenpmandclawhubtools to manage skill lifecycles, including installation, updates, and publication.\n- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by processing data from an external registry.\n - Ingestion points: Skill metadata, names, and descriptions are retrieved from
clawhub.comvia thesearchandinstallcommands.\n - Boundary markers: No delimiters or explicit instructions are used to distinguish remote registry content from system prompts.\n
- Capability inventory: The skill can execute shell commands (
npm,clawhub) and perform file system writes to the localskillsdirectory.\n - Sanitization: No sanitization or validation of the fetched registry content is described or performed before processing.
Audit Metadata