Get a 3-tier routing recommendation for a task: Tier 1 (Agent Booster, 0ms / $0 — for var-to-const, add-types, etc.), Tier 2 (Haiku — simple), Tier 3 (Sonnet/Opus — complex). Use this BEFORE spawning an agent to avoid sending simple transforms to Sonnet. Native tools have no equivalent — Claude C...
AI agents call hooks_route to retrieve information from Claude Flow without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool reads task characteristics and returns a routing recommendation (which tier/model to use). It does not spawn agents, execute code, modify data, or trigger any external operations. It is purely advisory/analytical, making it a Read category tool with low severity — at worst an incorrect recommendation could lead to suboptimal model selection, but it causes no direct harm.
From the tool's definition 'Get a 3-tier routing recommendation for a task' and 'Returns the recommended model' — the tool provides a recommendation/advisory output with no side effects
Attacks that exploit this kind of access
Get a 3-tier routing recommendation for a task: Tier 1 (Agent Booster, 0ms / $0 — for var-to-const, add-types, etc.), Tier 2 (Haiku — simple), Tier 3 (Sonnet/Opus — complex). Use this BEFORE spawning an agent to avoid sending simple transforms to Sonnet. Native tools have no equivalent — Claude Code does not introspect its own model-selection cost. Returns the recommended model + a. It is categorised as a Read tool in the Claude Flow MCP Server, which means it retrieves data without modifying state.
Register the Claude Flow MCP server in PolicyLayer and add a rule for hooks_route: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Claude Flow. Nothing to install.
hooks_route is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the hooks_route rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for hooks_route. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
hooks_route is provided by the Claude Flow MCP server (claude-flow). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.