The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx to fetch and run the @claude-flow/cli package from the public npm registry at runtime.
[REMOTE_CODE_EXECUTION]: By invoking npx @claude-flow/cli@latest, the skill executes code downloaded from a remote repository. The use of the @latest tag means the code executed is not pinned to a specific version, allowing for the execution of unvetted updates.
[PROMPT_INJECTION]: The skill implements real-time monitoring of agent and task events, which presents an attack surface for indirect prompt injection.
Ingestion points: Real-time event stream (NDJSON) generated by the swarm watch --stream command.
Boundary markers: There are no markers or system instructions to distinguish untrusted event data from authoritative instructions.
Capability inventory: The skill has the ability to execute shell commands (via Bash restricted to npx) and interact with MCP swarm status tools.
Sanitization: The skill lacks any sanitization or validation logic for the content of the streamed events.

monitor-stream