agent-hierarchy
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill implements a developer tool for repository management and agent coordination.
- [COMMAND_EXECUTION]: The
agent-guard.mjsscript utilizesexecFileSyncto run localgitcommands (ls-files,diff,status). These operations are necessary for auditing the repository's path ownership and ensuring agents stay within their defined surfaces. The commands are static and do not incorporate unvalidated external input. - [INDIRECT_PROMPT_INJECTION]: The skill defines a surface map (
docs/AGENT-SURFACES.md) which acts as a configuration source. While the agent-guard script parses this file, it is a local configuration file intended for use in developer environments and CI pipelines, posing minimal risk. The script implements simple regex-based parsing without executable evaluation of the content.
Audit Metadata