The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on the execution of the claude-flow CLI tool via npx for core functionalities, including initializing the collective memory system, spawning worker agents, and monitoring performance metrics.
[EXTERNAL_DOWNLOADS]: The use of npx implies the potential download and execution of the claude-flow package from the public npm registry during initialization or execution.
[INDIRECT_PROMPT_INJECTION]: The skill defines a system where a 'Queen' agent orchestrates tasks based on user-provided objectives. This creates a surface for indirect prompt injection if malicious instructions are embedded in the task objectives or memory entries that the agents process.
Ingestion points: Objective strings provided to the spawn command and data stored within the CollectiveMemory system (e.g., knowledge, task, and result types).
Boundary markers: The provided documentation does not define specific delimiters or instructions to treat objective inputs as untrusted data.
Capability inventory: The system has the capability to spawn new processes (worker agents), perform file system writes (SQLite database for memory), and integrate with external services like GitHub for repository analysis.
Sanitization: There is no evidence in the instructions of input validation, escaping, or sanitization for the objectives passed to the agents.

hive-mind-advanced