Skip to main content
Glama

submit_stage

Writes an external stage's answer into the run folder and resumes the run in the background once every waiting stage is answered.

Instructions

Write the answer for an external stage into the run folder, then resume the run in the background. The stage must be one external_prompt lists under waiting; the run resumes once all of them are answered. claimed_by is optional (v3 §1): a self-declared peer-session name, recorded as this stage's claim; every check here is warn-only and never blocks the write, so a caller that omits it sees exactly today's behavior. answered_model is optional: the model that wrote this answer, as you know it (for example claude-opus-5-5 or gpt-6.1). It is recorded as your own statement and is never checked; leave it out if you do not know.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "run": "<run>",
      "stage": "<stage>",
      "content": "<content>"
    },
    "name": "submit_stage"
  }
}

Input Schema

NameRequiredDescriptionDefault
runYes
stageYes
contentYes
claimed_byNoself-declared peer-session name, recorded as this stage's claim before the answer is written
session_readsNoOptional. Project files you read yourself to write this answer, as a list of { path, lines, why }: the path in the project, the lines you read as "3-5" (leave out for the whole file), and one sentence on what you used it for. The harness cuts those lines itself, scans them for keys, and shows the text to every later seat. A claim about code you read that is not listed here stays unverified.
answered_modelNothe model that wrote this answer, in your own words; recorded as declared, never verified
excerpt_answersNoOptional. Your answers to the text requests listed in this stage's prompt, as a list of { request, path, lines } to paste (request is the number in the list; path and lines default to what was asked) or { request, decline } to decline with one sentence. The harness cuts the lines itself and shows the text to every later seat.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed2 schema fields changedv0.8.4
    • addedInput schema / properties / excerpt_answers
      Added value: +{
      +  "description": "Optional. Your answers to the text requests listed in this stage's prompt, as a list of { request, path, lines } to paste (request is the number in the list; path and lines default to what was asked) or { request, decline } to decline with one sentence. The harness cuts the lines itself and shows the text to every later seat.",
      +  "items": {
      +    "properties": {
      +      "decline": {
      +        "type": "string"
      +      },
      +      "lines": {
      +        "type": "string"
      +      },
      +      "path": {
      +        "type": "string"
      +      },
      +      "request": {
      +        "maximum": 9007199254740991,
      +        "minimum": -9007199254740991,
      +        "type": "integer"
      +      }
      +    },
      +    "required": [
      +      "request"
      +    ],
      +    "type": "object"
      +  },
      +  "type": "array"
      +}
    • addedInput schema / properties / session_reads
      Added value: +{
      +  "description": "Optional. Project files you read yourself to write this answer, as a list of { path, lines, why }: the path in the project, the lines you read as \"3-5\" (leave out for the whole file), and one sentence on what you used it for. The harness cuts those lines itself, scans them for keys, and shows the text to every later seat. A claim about code you read that is not listed here stays unverified.",
      +  "items": {
      +    "properties": {
      +      "lines": {
      +        "type": "string"
      +      },
      +      "path": {
      +        "type": "string"
      +      },
      +      "why": {
      +        "type": "string"
      +      }
      +    },
      +    "required": [
      +      "path",
      +      "why"
      +    ],
      +    "type": "object"
      +  },
      +  "type": "array"
      +}
  2. Changed1 schema field changedv0.8.3
    • addedInput schema / properties / answered_model
      Added value: +{
      +  "description": "the model that wrote this answer, in your own words; recorded as declared, never verified",
      +  "maxLength": 200,
      +  "type": "string"
      +}
  3. First observedv0.8.1

TDQS

A4.3/5.0
Behavior4/5

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

With no annotations, the description carries the full burden and does substantial work: it discloses the background resume side effect, that all checks on `claimed_by` are warn-only and never block the write, that omitting it preserves current behavior, and that `answered_model` is recorded but never verified. It does not say what happens on a rejected write or what the call returns, leaving some behavioral gaps.

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 core behavior and the precondition are front-loaded in the first two sentences, and the optional-parameter notes follow logically. Some phrasing is redundant, such as the '(v3 §1)' citation and the restatement that omitting `claimed_by` yields today's behavior, which repeats the warn-only point.

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

Completeness4/5

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

For a seven-parameter mutation tool with no annotations and no output schema, the definition covers the write, the background resume, the `waiting`-stage precondition, and the trust level of the two declared fields. It leaves out error behavior and return shape, which an agent would need to handle failures confidently.

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 description coverage is 57%, and the description compensates by explaining the meaning of `claimed_by` (a self-declared peer-session name recorded as this stage's claim, with warn-only validation) and `answered_model` (recorded as the caller's own statement, never checked, with example values). It adds real semantics beyond the schema, though the three required parameters `run`, `stage`, and `content` receive no explanation in either place.

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 opens with a concrete verb-and-resource pair: it writes an external stage's answer into the run folder and then resumes the run in the background. It also differentiates itself from the sibling external_prompt by naming it as the source of the valid `waiting` stages, so an agent can tell which tool lists stages and which one answers them.

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?

It states a hard precondition ('The stage must be one external_prompt lists under `waiting`') and explains the completion semantics: the run resumes only once all waiting stages are answered. It gives no explicit when-not guidance or alternatives for edge cases (for example, what to do if the stage is not in `waiting`), which keeps it short of a 5.

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