Skip to main content
Glama

Proof Holdings

by holdings.proof
OverviewDiscussionsAdmin

create_confirmation

Create a HITL confirmation request. Sends an approval request to the channels configured on the referenced HITL config. First response wins.

Agent usage: The HITL config referenced by hitl_id must be authorized before you can create confirmations. If this returns a 403 or "not authorized" error, call request_hitl_authorization on the HITL config first, then wait for the user to approve via the messaging channel. After creating a confirmation, poll get_confirmation to check for approval/denial, or use get_confirmation_approval_link to generate a fresh link for the user.

Encryption is NOT optional: a HITL config cannot send confirmations without a keypair, the human approves by decrypting in their browser, and message must be a JSON ciphertext envelope. Plaintext is refused with 400 message_not_encrypted; an envelope sealed to another key with 400 message_key_mismatch. Choose the version by who must be able to decrypt:

  • v1 (single-recipient): wraps the content key for the config owner only. Use ONLY when there are no enrolled non-owner approvers — with approvers enrolled it is refused with 400 envelope_missing_approvers.

  • v2 (multi-recipient): additionally wraps the content key for each enrolled approver via an approver_ek map, so every approver can decrypt. Build it with the SDK encryptMessageV2(plaintext, ownerPublicKey, hitlId, approverKeys) after fetching the approvers via the SDK getApproverKeys(hitlId) (GET /api/v1/hitl/:id/approver-keys). AAD is the hitlId.

ACCESS: needs a Proof account. Authenticate this client (Claude Code: /mcp → Authenticate), then call this tool again. start_login does NOT open this tool.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
hitl_idYesHITL config ID whose channels will receive the approval request
messageYesThe client-encrypted ciphertext envelope as a JSON string (v1 or v2) — NOT the human-readable text. Encrypt the sentence the approver should read; passing it directly is refused with 400 `message_not_encrypted`, because the approval page can only decrypt an envelope.
proof_expiry_daysNoPer-request proof expiry override in days (7, 30, 90, or 365)

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed2 schema fields changed
    • changedInput schema / properties / message / description
      Previous value: -"Human-readable description of what needs approval"New value: +"The client-encrypted ciphertext envelope as a JSON string (v1 or v2) — NOT the human-readable text. Encrypt the sentence the approver should read; passing it directly is refused with 400 `message_not_encrypted`, because the approval page can only decrypt an envelope."
    • changedInput schema / properties / message / maxLength
      Previous value: -2000New value: +10000
  2. First observed

TDQS

A5/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full behavioral burden and meets it: it discloses destination channels, first-response-wins semantics, mandatory encryption, the 403/400 failure modes, and v1/v2 recipient behavior. The agent can predict outcomes and error handling before calling.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is long but modular: summary sentence, 'Agent usage', 'Encryption' with v1/v2 bullets, and 'ACCESS'. Every sentence earns its place given the encryption complexity and auth prerequisite, and critical information is front-loaded in the first sentence.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a state-changing tool with nuanced encryption requirements and no output schema, the description covers preconditions, post-call polling, error recovery, and authentication. An agent has enough information to invoke it successfully and handle expected failures.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Although schema coverage is 100%, the description adds substantial meaning: message must be a ciphertext envelope, plaintext gets rejected, v1/v2 key-wrapping differences, AAD as hitlId, and how to construct v2 via SDK calls. This is exactly the kind of parameter semantics an agent cannot infer from the schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The opening sentence states the precise action and object: 'Create a HITL confirmation request' and clarifies that it sends an approval request to channels on a referenced HITL config. This is specific enough to distinguish it from siblings like create_hitl and get_confirmation without needing to open their schemas.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The 'Agent usage' section is an explicit playbook: authorize the HITL config first, call request_hitl_authorization on a 403, then poll get_confirmation or use get_confirmation_approval_link. It even warns that start_login does NOT open this tool, removing ambiguity about authentication.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Proof Holdings toolsInspect
Try in Browser
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources