Skip to main content
Glama

Freigeben

confirm_business_change
Destructive

Commit exactly the change on a trusted human Owner card. App-only: models must never invoke this tool, construct proof or request hidden metadata in chat. A chat “yes” is not the card’s human confirmation; present the existing card instead.

Keep one card intact

The trusted UI returns that card’s kind and approvalToken together. For kind=goal, pass its proposalId: this covers stored goal and General/role-budget requests. For kind=policy, pass its exact signed candidate instead. Never combine two cards, supply both selectors, patch the displayed change or alter signed revision/deadline data.

Authority remains current

Requires policy:write and current Owner/module access; policies also need policy:read. Role-budget retirement additionally needs existing work:read/write, accounting:read and Workspace/Accounting write authority. Each workflow validates signature, tenant, principal, grant and expiry. Goal changes, budget archives and policies also validate revision.

Read the result without retrying

approved records a goal/budget commit; confirmed records a policy commit with revision/policyVersion. not_applied caused no new effect. not_verified or a lost reply may still mean a commit occurred. pending_owner, expired or not_found do not prove success.

Never retry confirmation. Read get_business_change(proposalId), then the goal/budget by resource ID. For policies use the separately authorized Policy connection to read get_policy(kind=policy,moduleKey,policyKey). Receipts do not prove current state, and current state alone cannot attribute this attempt’s effect.

Input Schema

NameRequiredDescriptionDefault
kindYesCard variant from hidden metadata. goal covers stored goal and General/role-budget proposals; policy covers exact policy replacements.
candidateNoExact candidate from the matching card's hidden metadata, including the change and expiry. The trusted UI copies every field unchanged; never reconstruct it from chat.
proposalIdNoGoal variant only: exact proposal ID from the same card as approvalToken.
approvalTokenYesOpaque signed proof from this same human card’s hidden metadata. The trusted UI returns it unchanged; never construct it or request it from chat.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed1 schema field changed
    • changedInput schema / properties / kind / description
      Previous value: -"Card variant from hidden metadata. goal covers stored goal/general-budget proposals; policy covers exact policy replacements."New value: +"Card variant from hidden metadata. goal covers stored goal and General/role-budget proposals; policy covers exact policy replacements."
  2. Changed4 schema fields changed
    • changedInput schema / properties / candidate / properties / expectedRevision / description
      Previous value: -"Current revision returned for this tenant/module/policy. A stale revision requires a fresh read and renewed intent."New value: +"Opaque hash of the current complete state from that result, not a proposed-value hash or numeric history version. Stale state requires a fresh read and renewed intent."
    • changedInput schema / properties / candidate / properties / moduleKey / description
      Previous value: -"Policy module key returned by the same supported read/preparation as policyKey and expectedRevision."New value: +"Policy module key from the same supported get_policy_write_options result as policyKey and expectedRevision; never a tenant selector."
    • changedInput schema / properties / candidate / properties / policyKey / description
      Previous value: -"Supported policy identifier from that result; copy it with its associated module key and revision."New value: +"Existing editable policy identifier from that same result; arbitrary new identities and templates are unsupported."
    • changedInput schema / properties / candidate / properties / value / description
      Previous value: -"Exact requested policy value validated by the supported contract; copy a preparation result unchanged when supplied."New value: +"Complete explicitly requested replacement, not a patch: omitted fields are not retained. Must satisfy the selected native contract; canonical JSON is limited to 12000 UTF-16 code units. Copy supplied exact continuation arguments unchanged."
  3. Changed1 schema field changed
    • addedInput schema / allOf
      Added value: +[
      +  {
      +    "if": {
      +      "properties": {
      +        "kind": {
      +          "const": "goal"
      +        }
      +      }
      +    },
      +    "then": {
      +      "not": {
      +        "required": [
      +          "candidate"
      +        ]
      +      },
      +      "required": [
      +        "proposalId"
      +      ]
      +    }
      +  },
      +  {
      +    "if": {
      +      "properties": {
      +        "kind": {
      +          "const": "policy"
      +        }
      +      }
      +    },
      +    "then": {
      +      "not": {
      +        "required": [
      +          "proposalId"
      +        ]
      +      },
      +      "required": [
      +        "candidate"
      +      ]
      +    }
      +  }
      +]
  4. Changed1 schema field changed
    • changedInput schema / properties / approvalToken / description
      Previous value: -"Opaque approval proof supplied in hidden metadata by the matching show tool. The trusted UI copies it unchanged; never construct it or request it from chat."New value: +"Opaque signed proof from this same human card’s hidden metadata. The trusted UI returns it unchanged; never construct it or request it from chat."
  5. Added

TDQS

A4.8/5.0
Behavior5/5

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

Annotations already flag destructive=true and idempotentHint=false, and the description reinforces this with "Never retry confirmation" plus a full result taxonomy (approved, confirmed, not_applied, not_verified, pending_owner, expired, not_found) including which outcomes do not prove success. It further discloses authorization requirements (policy:write, policy:read, work:read/write, accounting:read, Workspace/Accounting write) and the validations performed (signature, tenant, principal, grant, expiry, revision) — all context the annotations cannot convey.

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?

Front-loaded with purpose and prohibition, then organized under three headers (card integrity, authority, result reading), so scanning is cheap. It is long and repeats a few warnings already present in the schema property descriptions ("never reconstruct it from chat"), which keeps it just short of top marks.

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?

There is no output schema, and the description compensates by enumerating result states and how to verify each without retrying. For a destructive, non-idempotent commit with four parameters and conditional selectors, nothing an agent needs to avoid a harmful call is missing.

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

Parameters4/5

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 a decision rule the schema structure alone does not articulate: kind=goal takes proposalId, kind=policy takes the exact signed candidate, and the two must never be combined or the displayed change patched. This meaningfully reduces mis-selection risk, though the field-level constraints (revision hash, expiry format) remain with 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?

"Commit exactly the change on a trusted human Owner card" names a specific verb (commit) and resource (a pending change tied to an approval card), and the app-only clause immediately separates it from read/withdraw siblings such as get_business_change and withdraw_business_change. An agent can tell what this does and who is allowed to call it before opening the schema.

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?

Explicitly states the when-not ("models must never invoke this tool", "a chat 'yes' is not the card's human confirmation") and the alternative ("present the existing card instead"). It also routes post-call verification to named tools (get_business_change(proposalId), then goal/budget by resource ID, or get_policy for policies), so both the entry and follow-up paths are specified.

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

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources