harness-adapters
Pass
Audited by Gen Agent Trust Hub on Jul 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill functions as a documentation and reference resource for internal agent operations. It does not contain executable scripts or malicious instructions.
- [COMMAND_EXECUTION]: The skill documents the expected behavior and command strings for several third-party AI CLI tools. It references internal shell scripts (e.g.,
bin/fm-spawn.sh,bin/fm-harness.sh) used by the agent framework to automate these tools. This usage is consistent with the skill's stated purpose of harness management. - [SAFE]: The skill describes reading and writing configuration files within a controlled directory structure (
config/,state/, and.fm-grok-turnend). - [SAFE]: Interaction with the user's home directory is documented for integration purposes, specifically for installing global hooks for the Grok harness in
~/.grok/hooks/. This behavior is explicitly explained as a way to provide per-turn automation without modifying project-local files. - [SAFE]: The skill provides instructions on how the agent can automate responses to standard interactive CLI prompts, such as directory trust dialogs, by sending simulated keystrokes. This is a standard automation pattern for the described framework.
Audit Metadata