orca-workflow
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local orchestration commands (orca, agy, opencode) and standard Unix utilities (docker, printf, git). These are used for legitimate workflow management, status reporting, and safety checks.
- [EXTERNAL_DOWNLOADS]: The skill clones a repository from GitHub (github.com/rheinmir/setup) to synchronize its configuration. This repository belongs to the skill's author (rheinmir), which is consistent with the skill's intent.
- [DATA_EXFILTRATION]: The skill performs git push operations to the author's repository. This is used specifically to synchronize the skill's own code (SKILL.md) and does not involve exfiltrating sensitive user data or credentials.
- [PROMPT_INJECTION]: The workflow incorporates a mandatory 'gate' step requiring explicit user approval before implementation or dispatching tasks to other agents. This human-in-the-loop requirement provides a strong defense against automated prompt injection attacks.
Audit Metadata