The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to execute a local binary named anywrite. It requires the agent to run shell commands with user-provided absolute paths to perform operations on the Anytype API.
[PROMPT_INJECTION]: The skill retrieves and processes data from external sources via the Anytype local API, including object content, search results, and chat messages. This data is untrusted and could contain malicious instructions (indirect prompt injection).
Ingestion points: Data enters the agent's context through anywrite objects, anywrite search, and anywrite chat command outputs.
Boundary markers: There are no specified delimiters or instructions to ignore embedded commands within the retrieved Anytype content.
Capability inventory: The agent can execute shell commands via the anywrite binary and potentially other system tools.
Sanitization: No sanitization or filtering logic is described for the content retrieved from the API before it is processed by the agent.
[CREDENTIALS_UNSAFE]: The skill manages an API key stored in local configuration files at ~/.anywrite/config.json and ~/.anytype-cli/config.yaml. It also utilizes the ANYTYPE_API_KEY environment variable for authentication.

anywrite