ship-simple-change
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFE
Full Analysis
- [CREDENTIALS_UNSAFE]: The skill explicitly instructs the agent to scan staged diffs for hardcoded secrets, API keys, and credentials in Step 3, and to halt the process if any are discovered. This is a security-positive defensive check.
- [COMMAND_EXECUTION]: Utilizes standard development CLI tools including
git,npm, andghto perform common repository tasks (staging, linting, testing, committing, and PR creation). - [EXTERNAL_DOWNLOADS]: Performs standard
git fetchoperations from the configured origin to ensure the local environment is synchronized with the remote repository. - [DATA_EXFILTRATION]: Network activity is restricted to standard developer operations: pushing code to the origin repository and creating pull requests on GitHub. It also generates links to organizational Jira instances (
atlassian.netandguideline.tools) for issue tracking. - [PROMPT_INJECTION]: Instructions regarding "uninterrupted" execution are clearly scoped to user-requested workflow automation and do not attempt to bypass core AI safety filters or exfiltrate system prompts.
Audit Metadata