setup-conventions
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references configuration schemas and reusable CI workflows hosted at the author's GitHub repository (
github.com/anivar/agentic-sdlc). - [REMOTE_CODE_EXECUTION]: The GitLab CI template (
assets/gitlab-ci.yml) is designed to download and execute code directly from the vendor's repository using theuvxtool. - [COMMAND_EXECUTION]: Performs repository inspection and configuration using standard CLI tools including
gitand the GitHub CLI (gh). - [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: Reads existing repository state and files (e.g.,
AGENTS.md,git remote) to determine configuration parameters. - Boundary markers: Employs mandatory user confirmation rules (ASK-1, ASK-2, ASK-3) before performing writes, commits, or imports.
- Capability inventory: Includes file writing, git hook installation, and CI configuration.
- Sanitization: Relies on a human-in-the-loop "decider" to review all proposed changes and pull requests.
Audit Metadata