AI agents invoke claude_code_session to trigger actions in Claude Code. 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.
The tool can list/inspect sessions (Read), but also cancel or interrupt them (Execute/Destructive). Cancelling or interrupting an active autonomous coding session can halt in-progress work, potentially leaving code in an inconsistent state. Since the most severe applicable category wins, and cancellation/interruption is an active operation with side effects (not purely reading), Execute is appropriate.
From the tool's definition "cancel, or interrupt sessions" — terminating or interrupting active sessions is an irreversible side-effect operation that affects running autonomous coding tasks
Documented attack patterns abuse exactly the kind of access claude_code_session gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Claude Code, and nothing reaches the server without passing your rules. This is the rule we recommend for claude_code_session:
{
"version": "1",
"default": "deny",
"tools": {
"claude_code_session": {
"limits": [
{
"counter": "claude_code_session_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}claude_code_session stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
List, inspect, cancel, or interrupt sessions. It is categorised as a Execute tool in the Claude Code MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Claude Code MCP server in PolicyLayer and add a rule for claude_code_session: 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 Code. Nothing to install.
claude_code_session 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 claude_code_session 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 claude_code_session. 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.
claude_code_session is provided by the Claude Code MCP server (xihuai18/claude-code-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Claude Code, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
3 Claude Code tools catalogued and risk-classified — across an index of 43,000+ MCP servers.