The Agent Skills Directory

[SAFE]: The skill is entirely instructional, providing a conceptual framework for tool design (e.g., Abstraction Ladders, Task Bundles, Batch Operations). It does not contain any functional scripts, binaries, or command-line instructions.
[SAFE]: Analysis of the markdown content and YAML metadata shows no evidence of prompt injection, obfuscation, or data exfiltration attempts. The examples provided (such as execute_query or delete_files) are illustrative text within documentation and are not executable within the context of the skill itself.
[SAFE]: The skill lacks any network requests, file system modifications, or persistence mechanisms. It serves as a static reference for developers using the platform.

tool-design-composition