The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the sops CLI to decrypt files and jq to parse the resulting JSON data.
[CREDENTIALS_UNSAFE]: The skill targets sensitive files including .env.sops and directories such as secrets/. It includes detailed mitigation instructions to ensure decrypted credentials are not leaked into the agent's output, logs, or persistent documentation.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The agent is instructed to read and process content from encrypted repository files (e.g., components.sops.json, .env.sops).
Boundary markers: The skill does not provide specific delimiters for the decrypted data.
Capability inventory: The agent has access to shell execution (sops, jq) and file system operations.
Sanitization: The instructions recommend filtering data with jq to extract only necessary metadata (e.g., registry keys) rather than processing full decrypted blobs.

sops-secret-access