@qed-proof/mcp

Submit a claim for verification

submit_claim

Submit an agent-performed action for verification, returning a claim ID and signed receipt. Queued checks are retrievable later.

Instructions

Records that an agent says it performed an action, so QED Proof can check the destination and issue a signed receipt. Returns a claim_id and, when the check finishes inline, the verdict and receipt_id; otherwise state is "queued" and get_verdict returns the outcome later.

Documented actions (verifier profiles published in the QED Proof spec):

github.checks.pass: Every check run on the claimed commit completed without failing. target: owner/repo; params: sha (string, 40 lowercase hex)
github.commit.push: The claimed commit exists on the remote and the claimed branch contains it. target: owner/repo; params: sha (string, 40 lowercase hex), branch (string)
github.pr.open: The claimed pull request exists, optionally against the claimed base branch and at the claimed head commit. target: owner/repo; params: number (integer ≥ 1), base? (string), head_sha? (string)
slack.message.post: A message with exactly the claimed timestamp exists in the claimed channel, and (optionally) its text matches a target: slack:///, with a team ID T… and a public (C…) or private (G…) channel ID,; params: ts (string, ^[0-9]{10}\.[0-9]{6}$), text_sha256? (string, 64 lowercase hex)
x.post.publish: The claimed post exists on X, was written by the account the target handle is connected as, and (optionally) its text target: @handle (1–15 of A–Z a–z 0–9 _, with the @); params: post_id (string of 1–20 digits), text_sha256? (string, 64 lowercase hex)

Any other action string is accepted and decided as "unverifiable" (reason unsupported_action). target and params are copied into the receipt, which is public and permanent; params are limited to 4 KB. Resubmitting with the same client_claim_id returns the existing claim instead of creating a new one.

Input Schema

TableJSON Schema

Name	Required	Description
`action`	Yes	Action string, e.g. github.commit.push.
`params`	No	The fields the action's profile requires.
`target`	Yes	Where the work was meant to land, in the action's own format (e.g. owner/repo).
`agent_id`	No	The agent's own identifier. Defaults to the MCP client's name.
`claimed_at`	No	When the agent says it did the work (ISO 8601 with a timezone). Defaults to now.
`client_claim_id`	No	A stable id for this claim, so a retry doesn't create a duplicate. Generated when omitted.

Schema Changelog

Changes observed during successful MCP inspections.

First observedv0.1.0Sep 27, 2026

TDQS

A4.7/5.0

Behavior5/5

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

The description goes well beyond the annotations. It discloses idempotency behavior ('Resubmitting with the same client_claim_id returns the existing claim'), permanence of receipts ('target and params are copied into the receipt, which is public and permanent'), parameter size limits ('params are limited to 4 KB'), handling of unsupported actions ('decided as unverifiable (reason unsupported_action)'), and the conditional return of verdict inline versus queued. This is rich behavioral context that annotations alone (readOnlyHint=false, idempotentHint=false) do not provide.

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

Conciseness4/5

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

The description is long but every section earns its place. It opens with the core purpose, then explains the return flow, and devotes a structured list to documented actions. It could arguably be trimmed, but the length is justified by the need to cover multiple action profiles. Front-loading the purpose and key behaviors helps an agent quickly grasp the tool's role.

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 complex tool with 6 parameters, nested objects, no output schema, and multiple action profiles, the description is comprehensive. It explains return values (claim_id, verdict, receipt_id, state), idempotency, permanent receipt implications, unsupported action handling, and target/param formats. Nothing an agent needs to correctly invoke the tool is missing.

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?

Schema coverage is 100%, so the baseline is 3, but the description adds substantial meaning beyond the schema. It documents each supported action with exact target formats (e.g., slack://<team_id>/<channel_id>, @handle) and parameter constraints (sha as 40 lowercase hex, ts regex, text_sha256). It also explains the purpose of client_claim_id for idempotency and the 4 KB limit on params. This is far more than the schema provides.

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 description states a specific verb and resource: 'Records that an agent says it performed an action, so QED Proof can check the destination and issue a signed receipt.' It clearly distinguishes this from siblings like get_verdict (which retrieves outcomes later) and get_receipt (which presumably retrieves receipts). The purpose is unambiguous and action-oriented.

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

Usage Guidelines4/5

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

The description implies when to use this tool: when an agent has performed an action and wants verification. It explicitly notes that if the check does not finish inline, state is 'queued' and get_verdict returns the outcome later, which guides the agent to the sibling tool. However, it does not explicitly state 'use this tool instead of X' or list exclusions; the guidance is inferred from the flow description rather than stated as a rule.

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

Deploy Server

Submit a claim for verification

Instructions

Input Schema

Schema Changelog

TDQS

Other Tools