auto-dev

Pass

Audited by Gen Agent Trust Hub on Mar 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes multiple Git commands (e.g., git rev-parse, git diff, git status) to validate the repository state and ensure it is operating within the correct branch and worktree.
  • [COMMAND_EXECUTION]: It utilizes the GitHub CLI (gh) to perform authentication checks and trigger GitHub Actions workflows for the current branch.
  • [COMMAND_EXECUTION]: The auto-dev-deploy-dev.sh script executes a local 'inference script' (typically .skills-hub/auto-dev/infer-targets.sh) to dynamically calculate deployment parameters based on modified files. This is an intended extension mechanism for project-specific logic.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from the repository (file diffs, commit messages) and external sources (web browsing via Playwright/Chrome MCP).
  • Ingestion points: Git diff outputs, project-specific inference script outputs, and external web content from browser MCP tools.
  • Boundary markers: No explicit delimiters or 'ignore instructions' markers are implemented for the processed data.
  • Capability inventory: The skill possesses significant capabilities including shell command execution (bash), GitHub Actions management (gh), and cloud resource inspection (GCP/Firestore).
  • Sanitization: The deployment script performs basic parsing of the inference script's output, looking for specific workflow= and input: prefixes.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 18, 2026, 10:43 AM