The Agent Skills Directory

[COMMAND_EXECUTION]: The skill generates and executes shell commands using node and npx. It specifically calls a local script located at plugins/ruflo-cost-tracker/scripts/outcome.mjs and the @claude-flow/cli utility to process model outcomes.
[EXTERNAL_DOWNLOADS]: The skill uses npx to execute @claude-flow/cli@latest, which involves fetching the package from the npm registry at runtime.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: Reads usage and optimization data from the cost-tracking and cost-patterns namespaces using the memory_search tool.
Boundary markers: No explicit delimiters are used for the ingested memory data within the prompt instructions.
Capability inventory: Includes command execution (node, npx), file storage via memory tools, and network communication via the model-outcome hook.
Sanitization: The skill instructions explicitly note the use of spawnSync with explicit-argv in the outcome script to prevent command injection when handling task descriptions as arguments.

cost-optimize