The Agent Skills Directory

[COMMAND_EXECUTION]: The skill defines several commands to execute performance benchmarks using the 'npx agentic-flow workers benchmark' syntax. These commands target specific subsystems like 'trigger-detection', 'registry', and 'agent-selection'.
[EXTERNAL_DOWNLOADS]: The benchmark execution relies on 'npx', which downloads and runs the 'agentic-flow' package from the NPM registry. This is standard behavior for the specified tool and matches the skill's stated purpose.
[SAFE]: The skill focuses on metrics collection, latency testing, and optimization recommendations within its own ecosystem. The use of unconventional character separators (such as '$' in package names or metrics) appears to be a stylistic or platform-specific convention rather than an attempt at obfuscation.

worker-benchmarks