The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructions direct the agent to load API credentials from local configuration files at ~/.config/scry/env or the legacy path ~/.scry/.env. Accessing local files to retrieve secrets is a sensitive operation, although the skill follows standard secret management practices for CLI-based tools.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from a wide range of external, untrusted sources such as Reddit, Twitter, and web crawls. This creates an attack surface where instructions embedded in the retrieved data could potentially influence the agent's behavior.
Ingestion points: SQL query results from /v1/scry/query and passage quotes from the brief tool.
Boundary markers: While the skill emphasizes provenance and reporting the space covered, it lacks explicit boundary markers or instructions to the agent to disregard commands found within the retrieved data.
Capability inventory: The skill uses network operations to communicate with the Scry API and executes SQL statements via MCP tools.
Sanitization: No specific sanitization or filtering logic is provided to mitigate prompt injection risks in the processed data.
[COMMAND_EXECUTION]: The "Starter" section provides a shell script example that utilizes curl and environment loading. The agent may execute these or similar commands to interact with the service or verify the setup.
[EXTERNAL_DOWNLOADS]: The skill makes frequent network requests to api.scry.io and scry.io to retrieve schema information, examples, and query results. These domains are official resources for the vendor 'exopriors'.

scry