cutting-a-release
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute various shell commands to manage the git workflow, including branch switching, merging, and committing. It also executes a local synchronization script
bash scripts/common/sync-version.shto update version identifiers across the repository.- [EXTERNAL_DOWNLOADS]: The skill performs network operations viagit pullandgit fetchto synchronize the local environment with the remote origin repository.- [DATA_EXFILTRATION]: The skill usesgit pushto transmit local repository changes, including version files and generated documentation, to the remote server.- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it processes untrusted input from commit messages and code diffs to generate release notes. - Ingestion points:
SKILL.md(§4) directs the agent to read output fromgit logandgit showto identify changes. - Boundary markers: The instructions do not define delimiters or provide warnings to the agent to ignore instructions embedded within the git history data.
- Capability inventory: The agent has permissions to write to multiple project files and execute shell commands and repository scripts.
- Sanitization: There is no evidence of sanitization or filtering applied to the commit summaries or diffs before they are interpolated into the documentation generation process.
Audit Metadata