add-git-tag

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Executes shell commands via the Bash tool to interact with the Git repository.
  • Used to retrieve repository history, status, and tags in Steps 1, 3, and 6.
  • Used to create annotated tags (git tag -a) and push them to remote repositories (git push origin) in Steps 4 and 5.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface where untrusted data from the Git repository could influence agent behavior.
  • Ingestion points: Reads commit messages and repository metadata via git log --oneline --decorate -30 and git status --short in Step 1.
  • Boundary markers: Absent. The suggested sidecar prompt does not use delimiters to separate ingested git history from the instructions.
  • Capability inventory: The skill has access to Bash (allowing git modifications and network pushes), Write, and Read tools.
  • Sanitization: None. The skill does not sanitize or escape Git log output before passing it to the LLM (sidecar) for processing.
  • [DATA_EXFILTRATION]: Performs network operations by pushing tags to the remote Git repository (origin).
  • This is the intended primary function of the skill and is gated by explicit user confirmation in Step 5 before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 02:58 AM
Security Audit — agent-trust-hub — add-git-tag