The Agent Skills Directory

[SAFE]: The skill provides architectural patterns for tool concurrency control without any malicious instructions. It does not attempt to override system prompts or bypass safety guidelines.
[SAFE]: All code snippets are pedagogical examples for asynchronous programming in Python and do not include any remote code execution, obfuscation, or unauthorized access patterns.
[SAFE]: The skill proactively addresses security concerns related to shared state by recommending defensive strategies such as per-resource locking, mtime/hash verification for file edits, and workspace isolation for subagents.

concurrency-control