auto-dev-setup

Pass

Audited by Gen Agent Trust Hub on Oct 8, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill configures a workflow that triggers on external, untrusted input sources. Specifically:
  • Ingestion points: The workflow in SKILL.md (Step 3) is triggered by issues (labeled) and issue_comment (created) events.
  • Boundary markers: There are no specific instructions or boundary markers mentioned to prevent the agent from obeying instructions embedded within the issue body or comments.
  • Capability inventory: The stated purpose is "automated development," which typically involves file system modifications and execution of development tools.
  • Sanitization: The provided configuration does not include sanitization steps for the input strings (goal and adjustment) before they are passed to the development agent.
  • [COMMAND_EXECUTION]: The instructions include standard shell commands for environment setup:
  • mkdir -p .github/workflows and other directory creation commands are used to initialize the project structure.
  • [EXTERNAL_DOWNLOADS]: The skill references an external reusable workflow from the author's repository:
  • It suggests using uses: {SKILLS_REPO}/.github/workflows/auto-dev-reusable.yml@main for the GitHub Action. Documentation links confirm this points to the miles990/claude-software-skills repository, which belongs to the skill author.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 8, 2026, 09:40 PM