launch-openshell-gator
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a repository-local launcher script (
scripts/agents/run.sh) and various standard CLI utilities (git,gh,jq,ruby,openshell) for environment preparation, state inspection, and agent lifecycle management. - [DATA_EXFILTRATION]: Performs health checks on local authentication materials by verifying the presence of tokens in
~/.codex/auth.jsonand querying the official GitHub API viagh api user. These actions are limited to validating the execution environment and utilize well-known services corresponding to the skill's intended purpose. - [PROMPT_INJECTION]: Includes explicit 'Non-Negotiable Rules' and 'Input Normalization' sections that mandate the use of regular expressions to validate and sanitize user-supplied input (such as PR numbers, issue numbers, and gateway names) before they are incorporated into shell commands or generated prompts.
Audit Metadata