The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill explicitly references and utilizes the private SSH key path at ~/.ssh/id_ed25519 for decryption operations.
[CREDENTIALS_UNSAFE]: The skill integrates with the 1Password CLI (op) to retrieve, create, and process login credentials, which are then piped into other commands.
[COMMAND_EXECUTION]: The skill directs the agent to execute multiple shell commands including age, jq, op, git, and a deployment tool named hey.
[PROMPT_INJECTION]: The skill exhibits a vulnerability surface for indirect prompt injection because it processes content from external files (secrets.nix, .age files) and interpolates that data into shell commands or Nix configurations.
Ingestion points: Reads data from hosts/<host>/secrets/secrets.nix and decrypted content from .age files.
Boundary markers: Absent; the skill does not define delimiters or warnings to prevent the AI from following instructions embedded in secret values or configuration files.
Capability inventory: Extensive subprocess execution capabilities including age, op, git, and hey for deployment.
Sanitization: None; the skill uses direct shell piping (e.g., printf ... | age) which does not sanitize input for shell metacharacters or embedded instructions.

agenix-secrets