The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes system commands including ansible-playbook, git, brew, and npx to drive the verification process and check system configurations. Evidence: The verify.sh script wraps ansible-playbook and other CLI tools.
[CREDENTIALS_UNSAFE]: The skill accesses sensitive local file paths to verify identity and encryption roles within the playbook. Evidence: verify.sh references ~/.ansible-vault/vault.secret and features/git-config.md specifies verification of ~/.ssh/id_ed25519.pub.
[REMOTE_CODE_EXECUTION]: The skill invokes npx --yes skills update, which fetches and executes external code from the npm registry, and manages Homebrew package installations. Evidence: features/claude-config.md and features/homebrew-packages.md.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from the local repository to drive verification actions, creating a surface for indirect instructions if repository content is compromised. 1. Ingestion points: roles/ and features/ directories. 2. Boundary markers: Absent. 3. Capability inventory: includes ansible-playbook, npx, brew, git, find, ln, and rm. 4. Sanitization: Absent.

verify-dotfiles