Approve and EXECUTE a previously-issued Hands invocation by its single-use approval token. The token is returned by any confirm-required Hands tool; tokens expire after 60 seconds and CANNOT be reused. Side effect equals whatever the underlying Hand does — this can be highly destructive (running ...
AI agents invoke confirm_hand to trigger actions in Kontexta. 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.
This tool executes arbitrary actions (shell commands, file modifications) that were previously staged by a Hands tool. While the tool itself is a gating mechanism, confirming it triggers the underlying destructive operation.
From the tool's definition 'EXECUTE a previously-issued Hands invocation','Side effect equals whatever the underlying Hand does — this can be highly destructive (running arbitrary shell commands, modifying files, etc.)','The token IS the auth'
Attacks that exploit this kind of access
Approve and EXECUTE a previously-issued Hands invocation by its single-use approval token. The token is returned by any confirm-required Hands tool; tokens expire after 60 seconds and CANNOT be reused. Side effect equals whatever the underlying Hand does — this can be highly destructive (running arbitrary shell commands, modifying files, etc.), so only call when the user has authorised the pending action. The token IS the auth (no external auth, no rate limits). Invalid, expired, or already-consumed tokens return an inert text response, NOT an error. It is categorised as a Execute tool in the Kontexta MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Kontexta MCP server in PolicyLayer and add a rule for confirm_hand: 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 Kontexta. Nothing to install.
confirm_hand 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 confirm_hand 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 confirm_hand. 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.
confirm_hand is provided by the Kontexta MCP server (safiyu/kontexta). 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 →