Run Codex

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

The annotations already mark the tool as mutating, open-world, and destructive, so the bar is lower. The description adds useful behavioral context by explaining that Codex can read and edit files and run commands in the sandbox, and that long runs move to the background and return a job_id. It also discloses the return surface (final message, commands, thread_id), which goes beyond what annotations provide.

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 compact and front-loaded: it opens with the core purpose, then covers capabilities, return values, and backgrounding behavior in just a few sentences. Every clause earns its place and there is no filler or restating of the tool name.

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?

For a complex 16-parameter tool with no output schema, the description supplies the essential missing pieces: what the agent can do, what it returns, and how continuation and backgrounding work. Sandbox modes, timeouts, and safety flags are left to the fully covered schema, which is a reasonable division. A short explicit caution about destructive potential would push this to 5, but the destructiveHint annotation already covers that.

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?

Schema description coverage is 100%, so the schema already documents all 16 parameters and the baseline is 3. The description adds little parameter-level meaning beyond framing the prompt as the input to the agent and mentioning the thread_id/job_id outputs. It does not need to compensate, but it also does not enrich individual parameter understanding.

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 states a specific action ('Start a new Codex agent session') against a clearly identified resource ('a prompt'), and immediately distinguishes the tool from siblings by naming its two output tokens: thread_id (for codex_resume) and job_id (for background runs). This gives an agent a clear mental model of codex_exec as the entry-point tool rather than resume, review, or apply.

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?

The description clearly implies when to use the tool: for a brand-new session against a prompt. It also points to the follow-up tool by saying the returned thread_id can be passed to codex_resume, which helps an agent understand the workflow. It does not explicitly state when not to use it or name excluded siblings, so it falls just 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.