AI agents invoke session_wait to trigger actions in Pokeclaw. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
While session_wait itself is a polling operation, it is part of a suite of session control tools (session_start, session_send, session_kill, session_output) that collectively enable arbitrary code execution on the host machine. The tool allows monitoring and synchronization of that execution. This is Execute rather than Read because it participates in controlling active code sessions, not merely querying static data.
From the tool's definition Tool enables waiting for session output/exit on active Codex CLI or Claude Code sessions spawned on the host machine.
Attacks that exploit this kind of access
Wait for a session to produce new output or exit (polls up to timeout). It is categorised as a Execute tool in the Pokeclaw MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Pokeclaw MCP server in PolicyLayer and add a rule for session_wait: 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 Pokeclaw. Nothing to install.
session_wait is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the session_wait 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 session_wait. 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.
session_wait is provided by the Pokeclaw MCP server (madebydia/pokeclaw). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →