release-orchestrator
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill scripts (changelog_generator.py, preflight_checker.py, version_bumper.py) utilize subprocess.run to execute local git commands. While these are necessary for the skill's primary functionality—such as reading logs, checking branch status, and fetching updates—they represent a direct capability to interact with the host system's command line.
- [INDIRECT_PROMPT_INJECTION]: The skill parses and processes untrusted data from the repository, specifically Git commit messages for changelog generation and file contents for secret scanning. This creates a standard attack surface where malicious data in the repository could influence agent behavior.
- Ingestion points: Repository commit history (via git log) parsed in changelog_generator.py and various tracked files scanned in preflight_checker.py.
- Boundary markers: None; repository data is processed and interpolated into markdown reports or manifest files without explicit delimiters or warnings to ignore embedded instructions.
- Capability inventory: version_bumper.py can write to project manifest files (e.g., package.json, pyproject.toml), and multiple scripts execute git commands via subprocess.
- Sanitization: No evidence of escaping, filtering, or sanitizing external repository content was detected before the data is processed or outputted.
Audit Metadata