Start local Codex turn

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

Beyond the annotations (destructiveHint, openWorldHint), the description discloses important behavioral traits: the turn is persistent and runs asynchronously, and it is a high-privilege entry point that may access paths outside the structured root. This adds value over the structured metadata, making the agent aware of long-running side effects and security implications.

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

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

The description is exceptionally well-structured: three sentences, with the primary action and return immediately stated first, followed by the crucial async behavior and security warning. Every sentence carries significant information without fluff, making it both concise and effective.

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

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

Given the tool's complexity — 8 parameters, async behavior, security implications, and an output schema — the description covers only the most essential points (async, polling, thread creation, privilege level). It omits guidance on the remaining parameters, error handling, and does not describe the return format beyond 'IDs'. While the output schema exists and might fill some gaps, the lack of parameter documentation makes the description incomplete for correct usage.

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

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

With 0% schema description coverage, the description is the only source of parameter meaning. It only explains thread_id ('Omit thread_id to create a new local Codex thread') and implicitly mentions polling. The other seven parameters (model, effort, attempt, project, step_id, workflow_id) are left entirely unexplained, forcing the agent to guess their purpose. This is insufficient for a tool with 8 parameters.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the action ('Start a persistent app-server turn') and the immediate outcome ('return its IDs'). It distinguishes itself from job-based siblings by explicitly naming it a 'local Codex turn' and referencing persistent async behavior, though it does not directly compare to tools like start_codex_job. The core purpose is unambiguous.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

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

It provides concrete usage context: the turn is asynchronous and continues after the call returns, with an explicit instruction to poll via get_codex_turn_status. It also explains how to create a new thread vs. continue an existing one by omitting or providing thread_id. While it does not mention when not to use it or name explicit alternatives, the given guidance is actionable and sufficient for correct invocation in most cases.

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