adopt-standards

Pass

Audited by Gen Agent Trust Hub on Aug 31, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The scripts/select-release script fetches release information and source code from the author's GitHub repository. It resolves the latest version from https://github.com/lutzseverino/repository-standards/releases/latest and clones the repository from https://github.com/lutzseverino/repository-standards.git using git clone.
  • [COMMAND_EXECUTION]: The skill uses shell commands to manage and verify the downloaded repository. It executes git clone to fetch the specific release branch, git describe --tags --exact-match HEAD to verify the checkout matches the expected version tag, and git status --porcelain=v1 to ensure the downloaded checkout is clean and unmodified.
  • [INDIRECT_PROMPT_INJECTION]: The skill acts as a bootstrap that instructs the agent to ingest and follow new instructions from the downloaded release.
  • Ingestion points: The agent is directed to read the SKILL.md file found within the downloaded checkout at .agents/skills/adopt-standards/SKILL.md.
  • Boundary markers: Absent; the skill relies on the integrity of the source repository and version pinning.
  • Capability inventory: The agent is instructed to "Read and follow" the new skill and invoke its adapter with specific environment variables (REPOSITORY_STANDARDS_CHECKOUT).
  • Sanitization: The bootstrap script performs extensive validation, including checking that the checkout is clean, verifying the Git tag matches the requested version, and ensuring an internal VERSION file matches the expected metadata.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 31, 2026, 09:26 AM
Security Audit — agent-trust-hub — adopt-standards