update-installed-skills
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
npx ai-agent-skillscommand line tool to manage skill installations, including listing, checking for updates, and syncing content. - [EXTERNAL_DOWNLOADS]: Performs network operations to check for updates and download skill content from remote sources like GitHub during the
checkandsyncphases. - [INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface for indirect prompt injection through managed content. 1. Ingestion points: Remote repositories fetched via the
synccommand. 2. Boundary markers: Absent in the automated synchronization process. 3. Capability inventory: File system write access (overwritingSKILL.mdfiles). 4. Sanitization: The risk is mitigated by explicit instructions for the agent and user to perform a--dry-runand manual review before applying updates.
Audit Metadata