Capable

Commit reviewed updates

commit_reviewed_updates

DestructiveIdempotent

Commit the exact approved subset from review_proposed_updates through the base MCP contract. Pass the unchanged short-lived receipt, only proposal IDs the user explicitly approved, confirm:true, and a fresh UUID idempotency key (normally the returned review_id). The first confirmed call freezes the selection; retries with the same receipt and key return stored per-row outcomes without executing again. Receipt expiry blocks only a first commit; an exact retained replay remains available for the bounded ledger-retention window. If an expired review has no retained exact match, inspect the CRM before any fresh review. Each selected target write re-enters its complete Capable guard wrapper, so current policy, object, enum, role/scope, audit, telemetry, invalidation, and automation behavior still applies. A receipt proves the reviewed arguments are unchanged; each target tool still reads current CRM state at commit time. If a replay reports execution_in_progress, poll this exact receipt/key/selection again. If it reports execution_state_unknown, inspect CRM state first and never retry automatically or start a fresh review until the result is known; check its ledger only by replaying that exact receipt, key, and full selection.

When to use: After review_proposed_updates, when the user has explicitly selected the returned proposal IDs. Pass its unchanged receipt, those IDs, confirm:true, and a fresh UUID idempotency key — normally the returned review_id; reuse that exact UUID and full selection for retries so they replay the stored outcome without writing twice. Receipt expiry blocks a first commit, not an exact retained replay; if no exact retained match exists, inspect the CRM before any fresh review.

Example: Approve proposal IDs from that review and commit them.

Input Schema

TableJSON Schema

Name	Required	Description
`confirm`	No	Must be true to execute; omitted or false commits nothing and returns a prompt to get the user's explicit approval first.
`receipt`	Yes	The receipt string returned by review_proposed_updates, passed back unchanged (it is bound to that review, this workspace and user, and expires).
`idempotency_key`	Yes	A UUID minted once per review (normally the returned review_id); reuse the same key to replay stored outcomes instead of executing again.
`selected_proposal_ids`	Yes	The review's proposal ids (its proposals[].id) the user explicitly approved — unique, each present in the receipt; the first confirmed call freezes this set.

Output Schema

TableJSON Schema

Name	Required	Description	Default
`note`	No
`summary`	No
`outcomes`	No
`replayed`	No
`confirmed`	No
`review_id`	No
`review_commit`	No
`request_status`	No
`selection_mismatch`	No
`frozen_selected_proposal_ids`	No

Schema Changelog

Changes observed during successful MCP inspections.

First observedOct 6, 2026

TDQS

A4.7/5.0

Behavior5/5

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

Goes well beyond the annotations: it explains first-call freezing of the selection, replay returning stored per-row outcomes, ledger-retention bounds on retained replays, re-entry through full guard wrappers, and current CRM state being read at commit time. It adds explicit handling for execution_in_progress (poll) and execution_state_unknown (inspect CRM, never auto-retry), which no annotation conveys.

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

Conciseness3/5

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

Front-loaded and organized, but the 'When to use' paragraph largely restates the first paragraph's receipt-expiry, idempotency-key, and retry guidance verbatim, which is wasted space in an otherwise dense block of text.

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?

An output schema exists, so return values need no explanation, yet the description still covers the non-obvious terminal and retry states an agent must act on. Given the destructive, idempotent nature of the commit, nothing needed to call it safely 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 already 100%, so the baseline is 3, but the description adds operational meaning: the idempotency key is normally the returned review_id and must be reused with the full selection for replay, confirm:true is the approval gate, and the selection is frozen by the first confirmed call.

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?

States a specific verb (commit) and resource (the exact approved subset from review_proposed_updates) and immediately names the sibling that produces the input (review_proposed_updates). An agent can distinguish this from review_proposed_updates/propose_updates without opening any 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?

Explicit 'When to use' paragraph gives the trigger (after review_proposed_updates, user has selected proposal IDs), the required arguments, and the retry rule (reuse the exact UUID and full selection). It also states when-not paths: receipt expiry blocks only a first commit, and no fresh review until an unknown result is resolved.

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

All Capable tools Inspect

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

Need Help?