syncing-awesome-skills
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The
registry.mjsscript performs shell command execution usingexecFileSyncto handle Git operations (such as cloning, fetching, and worktree management) and to interact with the GitHub CLI (gh) for pull request creation. - [DYNAMIC_EXECUTION]: The skill workflow involves executing package manager commands (
npm install,pnpm install) and running arbitrary scripts (lint,test,typecheck) defined in thepackage.jsonof the repositories being synchronized. - [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, including
skills-lock.jsonandSKILL.mdfiles in the target repositories. This presents an indirect prompt injection surface where instructions could be embedded to influence the agent's behavior. 1. Ingestion points: Reads configuration and instruction files from both the source and target repositories. 2. Boundary markers: No explicit delimiters or instructions to ignore embedded content are used when processing these files. 3. Capability inventory: The skill can write to the file system, execute Git/GitHub commands, and run repository-defined shell scripts. 4. Sanitization: The skill does not perform validation or sanitization of the content fetched from synchronized repositories. - [EXTERNAL_DOWNLOADS]: The skill clones and maintains a local cache of skills from a repository on GitHub to facilitate the synchronization process.
Audit Metadata