Skip to main content
Glama
parisbs

codex-subagent-mcp

by parisbs
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Local

Continue a Codex session

codex_follow_up

Send a follow-up message to a previous delegation using its thread_id. Codex retains the earlier context, so only the new instruction needs to be sent. Like a delegation, it is sent to OpenAI and spends the user's Codex usage.

Instructions

Send a follow-up message to a previous delegation using its thread_id. Codex retains the earlier context, so only the new instruction needs to be sent. Like a delegation, it is sent to OpenAI and spends the user's Codex usage.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
labelNoYour own short name for this kind of task, stored verbatim in the local usage log and never sent to Codex: 1 to 64 characters, no control or bidirectional-control characters. A follow-up without one keeps its thread's label.
modelNoOverride the model for this turn. Defaults to the thread's last model from memory or, on a registry miss, Codex's session file; if neither has it, the configured default, else the call is refused.
promptYesThe follow-up instruction.
sandboxNoSandbox policy for this turn. Uses the configured default when omitted; read-only when unset.
thread_idYesThe thread_id reported by a previous codex_delegate call.
working_dirNoAbsolute directory to resume in. Defaults to the directory the thread last ran in.
auto_approveNoNot supported on follow-ups: true is refused and nothing runs.
output_schemaNoA JSON Schema object for this turn's final message, which then comes back as JSON in a delimited block. OpenAI's structured outputs apply: every object needs "additionalProperties": false and every property listed in "required". At most 64 KiB serialised.
timeout_secondsNo
reasoning_effortNoOverride the reasoning effort for this turn. Defaults to the thread's last effort when the model is unchanged, otherwise to the configured or model default.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed1 schema field changedv0.6.0
    • addedInput schema / properties / label
      Added value: +{
+  "description": "Your own short name for this kind of task, stored verbatim in the local usage log and never sent to Codex: 1 to 64 characters, no control or bidirectional-control characters. A follow-up without one keeps its thread's label.",
+  "type": "string"
+}
  2. Changed1 schema field changedv0.4.0
    • addedInput schema / properties / output_schema
      Added value: +{
+  "additionalProperties": {},
+  "description": "A JSON Schema object for this turn's final message, which then comes back as JSON in a delimited block. OpenAI's structured outputs apply: every object needs \"additionalProperties\": false and every property listed in \"required\". At most 64 KiB serialised.",
+  "propertyNames": {
+    "type": "string"
+  },
+  "type": "object"
+}
  3. Changed2 schema fields changedv0.3.0
    • changedInput schema / properties / model / description
      Previous value: -"Override the model for this turn. Defaults to the model the thread last ran with on this server; for a thread this server has no record of, the configured default, else the call is refused."New value: +"Override the model for this turn. Defaults to the thread's last model from memory or, on a registry miss, Codex's session file; if neither has it, the configured default, else the call is refused."
    • changedInput schema / properties / sandbox / description
      Previous value: -"Sandbox policy for this turn. Defaults to read-only."New value: +"Sandbox policy for this turn. Uses the configured default when omitted; read-only when unset."
  4. Changed4 schema fields changedv0.2.0
    • addedInput schema / properties / auto_approve / description
      Added value: +"Not supported on follow-ups: true is refused and nothing runs."
    • changedInput schema / properties / model / description
      Previous value: -"Override the model for this turn."New value: +"Override the model for this turn. Defaults to the model the thread last ran with on this server; for a thread this server has no record of, the configured default, else the call is refused."
    • changedInput schema / properties / reasoning_effort / description
      Previous value: -"Override the reasoning effort for this turn."New value: +"Override the reasoning effort for this turn. Defaults to the thread's last effort when the model is unchanged, otherwise to the configured or model default."
    • addedInput schema / properties / working_dir / description
      Added value: +"Absolute directory to resume in. Defaults to the directory the thread last ran in."
  5. Changed1 schema field changed
    • addedInput schema / properties / thread_id / pattern
      Added value: +"^[A-Za-z0-9][A-Za-z0-9_-]{0,199}$"
  6. First observedv0.1.0

TDQS

A4.1/5.0
Behavior4/5

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

Annotations already flag openWorldHint and readOnlyHint=false; the description adds genuinely non-structured behavior: Codex retains prior context, the call leaves the local machine for OpenAI, and it consumes the user's Codex usage quota. It stops short of saying whether the call blocks or hands back a job handle.

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

Conciseness5/5

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

Three tight sentences: mechanism first, then the context-retention insight that explains why the prompt can be short, then the cost warning. Nothing is padded.

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

Completeness3/5

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

With 10 parameters, no output schema, and an open-world external dependency, the description should say what comes back — a result, a job id for codex_job_status, or a streamed block (the output_schema parameter implies JSON returns but this isn't stated). Cost and context are covered; return and blocking behavior are not.

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

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 90%, so parameters like thread_id, model, sandbox and output_schema are already documented. The description only reinforces thread_id's provenance and adds nothing about the optional overrides — baseline 3 applies.

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 and resource — send a follow-up message to a previous delegation — and pins it to a thread_id, which immediately separates it from codex_delegate (new task) and the codex_job_* tools (polling). An agent can pick it without 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 Guidelines4/5

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

Makes the precondition explicit: you must already have a thread_id from a prior delegation, and only the new instruction needs sending. It does not say when to abandon the thread and start a fresh codex_delegate instead, so there is no exclusion guidance.

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

Deploy Server

Other Tools