git

Pass

Audited by Gen Agent Trust Hub on Aug 1, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to execute standard Git commands for repository management, including git stash, git fetch, git rebase, and git push. These operations are essential for the skill's stated purpose of managing Git workflows.
  • [DATA_EXFILTRATION]: The /git publish command transmits local code and commit history to remote repositories via git push and creates pull requests using the GitHub MCP tool. This is the intended behavior for a version control automation tool.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the local repository, specifically CI/CD configuration files and commit logs, to automate branch naming and PR generation.
  • Ingestion points: Scans CI/CD files (e.g., .github/workflows/*.yml, .gitlab-ci.yml) and git log output to extract branch names and commit types.
  • Boundary markers: The skill utilizes specific parsing rules (e.g., 'Extract only literal branch names', 'Scan commit prefixes') to delimit the data it processes.
  • Capability inventory: The agent has access to Bash for shell operations, Write for updating local project configuration (.fullstack-dev/config.json), and GitHub API tools for PR creation.
  • Sanitization: The instructions mandate filtering for specific literal strings and conventional commit formats, reducing the risk of the agent interpreting data as instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 1, 2026, 03:45 AM
Security Audit — agent-trust-hub — git