Skip to main content
Glama
Sign In

OrgGen AI

by ai.orggen
OverviewDiscussionsAdmin

set_project_metadata

Destructive

Writes a CUSTOM project metadata entry, under a key of your choice — meant for agent CHECKPOINTS ("where I left off") and any working information you want to find again later. The value is free-form JSON and REPLACES the key's previous value entirely (no merge); the write is ATOMIC, so two agents cannot create two concurrent entries under the same key. These values live OUTSIDE the schema: they are not in the version snapshots (restore_version does NOT restore them), nor in the attestation of a signed version, nor in the public/anonymous links. ⚠️ NOT TO BE CONFUSED with update_meta (title/subtitle/currency of the schema, VISIBLE on the canvas) nor with update_kind_metadata (the attributes of ONE block). ⚠️ The key config is RESERVED: the project configuration is written from the site by a governing role, and not by a tool — the call is REFUSED whatever your role (the configuration is read with get_config).

Input Schema

NameRequiredDescriptionDefault
keyYesIdentifier of the entry, in lowercase: a-z, 0-9, then ".", "_", "-", ":" (64 characters max). E.g. "checkpoint", "run:2026-08-18", "notes.migration".
valueYesThe JSON value to write — object, array, string, number, boolean or null. REPLACES the previous value of this key entirely (no merge). 32 KiB of JSON at most: beyond that the call is REFUSED, naming the measured size, without truncation.
expectedProjectIdNoUUID of the project targeted by THIS call. ACTUALLY targets this project instead of the session project. Refused without changing anything if access is not granted.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed3 schema fields changed
    • changedInput schema / properties / expectedProjectId / description
      Previous value: -"UUID du projet visé par CET appel. Cible RÉELLEMENT ce projet au lieu du projet de session. Refusé sans rien modifier si l'accès n'est pas accordé."New value: +"UUID of the project targeted by THIS call. ACTUALLY targets this project instead of the session project. Refused without changing anything if access is not granted."
    • changedInput schema / properties / key / description
      Previous value: -"Identifiant de l’entrée, en minuscules : a-z, 0-9, puis « . », « _ », « - », « : » (64 caractères max). Ex. « checkpoint », « run:2026-08-18 », « notes.migration »."New value: +"Identifier of the entry, in lowercase: a-z, 0-9, then \".\", \"_\", \"-\", \":\" (64 characters max). E.g. \"checkpoint\", \"run:2026-08-18\", \"notes.migration\"."
    • changedInput schema / properties / value / description
      Previous value: -"La valeur JSON à écrire — objet, tableau, chaîne, nombre, booléen ou null. REMPLACE intégralement la valeur précédente de cette clé (jamais une fusion). 32 Kio de JSON au maximum : au-delà l’appel est REFUSÉ en nommant la taille mesurée, jamais tronqué."New value: +"The JSON value to write — object, array, string, number, boolean or null. REPLACES the previous value of this key entirely (no merge). 32 KiB of JSON at most: beyond that the call is REFUSED, naming the measured size, without truncation."
  2. First observed

TDQS

A4.8/5.0
Behavior5/5

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

Annotations already declare destructiveHint=true, and the description goes well beyond that: the write REPLACES the key entirely with no merge, it is ATOMIC against concurrent writers under the same key, and the data is excluded from version snapshots, signed attestations, and public links. The reserved-key refusal 'whatever your role' is a concrete failure mode an agent would otherwise trip over.

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?

Purpose and intended use are front-loaded, followed by write semantics, then scope exclusions, then the reserved key — a sensible ordering. It is long and leans on caps and warning glyphs, with the no-merge/atomic points echoed in both the description and the `value` schema, but essentially every sentence carries information.

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 3-parameter mutation with no output schema, the description covers what gets written, what gets lost, concurrency behavior, visibility/exclusion from snapshots and public links, and the one key that will be refused. Nothing an agent needs in order 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 100%, so key/value/expectedProjectId are already documented, but the description adds a constraint absent from the schema: the key `config` is reserved and the call is refused regardless of role. The 32 KiB limit and no-merge semantics of `value` are restated from the schema rather than added.

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 first sentence names a specific verb and resource (writes a CUSTOM project metadata entry) and states the intended use case (agent checkpoints / working information). It explicitly separates itself from update_meta and update_kind_metadata, so an agent can route 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?

It gives a positive reason to use it ('agent CHECKPOINTS', working data you want to find later) and two explicit exclusions with the correct sibling named for each (update_meta for schema title/subtitle/currency, update_kind_metadata for block attributes). The reserved `config` key is also flagged as uncallable, with get_config named as the read path.

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

All OrgGen AI 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