release
Pass
Audited by Gen Agent Trust Hub on May 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill facilitates release automation by executing
gitandgh(GitHub CLI) commands. These commands are used to manage version tags, commit changelog updates, and create releases on GitHub. - [DATA_EXFILTRATION]: The skill uses
git pushandgh release createto upload changelog data and release metadata to GitHub's servers. This is standard and expected behavior for a release management tool. - [PROMPT_INJECTION]: The skill processes content from local markdown files (
TASKS.mdand task documents) to generate release notes and shell command arguments, creating a surface for indirect prompt injection. - Ingestion points: Reads task descriptions and PR references from
TASKS.mdand related task files. - Boundary markers: None; the skill parses markdown structures directly without specific delimiters or safety instructions for the ingested content.
- Capability inventory: The skill generates and executes shell commands (
git commit,git tag,gh release create) that incorporate the ingested data. - Sanitization: None; ingested descriptions are interpolated directly into command templates and markdown files.
Audit Metadata