The Agent Skills Directory

[SAFE]: The skill implements a structured workflow for resolving uncertainty, using a destination-driven map and discrete investigation records. Its operations are purely procedural and do not contain malicious instructions.
[SAFE]: External resource interaction is conducted through well-defined contracts (e.g., and-backend-contract, and-interview-contract) and specific tools like research and prototype. These tools are integrated into the workflow suite and do not represent unauthorized remote code execution.
[SAFE]: The skill maintains project state using local configuration (.and/config.yml) and links to external gists for decision records. These are standard practices for development-oriented agents and utilize well-known services for storage.
[SAFE]: No obfuscation, such as Base64-encoded strings, zero-width characters, or hidden URLs, was found in the skill content.
[SAFE]: Security boundaries are explicitly defined, such as requiring dedicated investigation branches for prototype/research work and restricting delivery actions (e.g., costs, secrets, permissions) to human actors.

and-wayfind