Skip to main content
Glama
WebisityStudio

claude-codex-mcp-bridge

by WebisityStudio
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Local

Start an autonomous Codex implementation run

bridge_orchestrate_codex

Launches a saved Codex worker in an isolated, network-off git worktree and waits for its structured result, so you can delegate implementation or verification tasks and retrieve outcomes.

Instructions

Start a saved Codex worker in an isolated git worktree (network off) and wait up to waitSeconds for its structured result. If status is waiting_for_fable, answer the question and call bridge_continue_codex. If it is running_codex, the result arrives in the coordinator's mailbox. Independent calls create separate runs. This tool never grants permission to commit, push, deploy, publish, send externally, change credentials, delete data, or mutate production.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
taskYesSelf-contained implementation or verification task.
threadIdYesStable task/thread identifier.
maxRoundsNoMaximum coordinator/Codex turns. Defaults 6.
projectPathNoAbsolute path to the git repository. Defaults to this session's project directory.
useWorktreeNoCreate an isolated git worktree. Defaults true.
waitSecondsNoHow long this call waits for Codex. Defaults to 240 seconds. If Codex is still working, the result arrives in the coordinator's mailbox.
coordinatorAgentNoAgent that receives the result. Defaults to this conversation's registered agent, else claude-main.
includeUncommittedNoCopy the main checkout's uncommitted changes into the worktree.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed6 schema fields changedv0.4.0
    • changedInput schema / properties / coordinatorAgent / description
      Previous value: -"Name of the active Claude coordinator."New value: +"Agent that receives the result. Defaults to this conversation's registered agent, else claude-main."
    • addedInput schema / properties / includeUncommitted
      Added value: +{
+  "description": "Copy the main checkout's uncommitted changes into the worktree.",
+  "type": "boolean"
+}
    • changedInput schema / properties / maxRounds / description
      Previous value: -"Maximum Fable/Codex turns. Defaults 6."New value: +"Maximum coordinator/Codex turns. Defaults 6."
    • changedInput schema / properties / projectPath / description
      Previous value: -"Absolute path to the git repository."New value: +"Absolute path to the git repository. Defaults to this session's project directory."
    • addedInput schema / properties / waitSeconds
      Added value: +{
+  "description": "How long this call waits for Codex. Defaults to 240 seconds. If Codex is still working, the result arrives in the coordinator's mailbox.",
+  "maximum": 285,
+  "minimum": 0,
+  "type": "integer"
+}
    • changedInput schema / required
      Previous value: -[
-  "coordinatorAgent",
-  "projectPath",
-  "task",
-  "threadId"
-]New value: +[
+  "task",
+  "threadId"
+]
  2. First observedv0.1.0

TDQS

A4.3/5.0
Behavior5/5

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

With no annotations, the description carries the full burden and does so well: network off, isolated worktree, timestamped wait limit with mailbox fallback on timeout, and independent-call semantics. It also enumerates the actions the tool never authorizes (commit, push, deploy, publish, external send, credential change, data deletion, production mutation), which is exactly the safety context annotations would otherwise provide.

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?

Four sentences with the core purpose front-loaded, followed by state routing and the safety list. The permission enumeration is long but earns its place given absent annotations; otherwise there is no padding.

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 an 8-parameter orchestration tool with no annotations and no output schema, the description covers the operational states and the authorization boundary an agent most needs. The remaining gap is that 'structured result' is never characterized, and since there is no output schema the caller must infer the return shape from the mailbox/continuation flow.

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 100%, so all 8 parameters are already documented in the schema, making 3 the baseline. The description reinforces waitSeconds timeout semantics and the mailbox fallback, but adds no format, default, or constraint detail beyond what the schema already states.

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: 'Start a saved Codex worker in an isolated git worktree (network off)' and waits up to waitSeconds for its structured result. It also distinguishes itself from the follow-up sibling by naming bridge_continue_codex, so an agent can separate it from ask_codex/review_with_codex by its 'saved worker + worktree' framing.

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?

Gives explicit conditional routing: waiting_for_fable -> answer then call bridge_continue_codex; running_codex -> result arrives in the coordinator's mailbox. It also notes independent calls create separate runs. It does not, however, say when to prefer this over ask_codex or review_with_codex, so the alternative-selection guidance is incomplete.

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