agent-skill-update-monitor
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches skill updates and repository metadata from the vendor's official GitHub repository (
github.com/lensetek/Book-Author-Agent-Skills). - [COMMAND_EXECUTION]: Executes shell commands such as
git fetchandgit cloneto interact with the repository and update local files. - [DATA_EXFILTRATION]: Accesses local project directories and
SKILL.mdfiles to compare installed versions with remote repository versions. - [PROMPT_INJECTION]: Processes content from external repositories to generate update summaries, which presents a surface for indirect prompt injection.
- Ingestion points: Data enters the agent context from the remote GitHub repository (e.g.,
SKILL.mdfiles, commit messages, and README content). - Boundary markers: Absent; remote data is processed and summarized without explicit delimiters or boundary markers defined in the instructions.
- Capability inventory: The skill has capabilities to perform network requests to GitHub and file system writes to update local skill files.
- Sanitization: Instructions explicitly mandate treating downloaded files as untrusted and recommend running a security/privacy checker before final installation.
Audit Metadata