Delegate in background (paid)

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

Beyond the annotations (readOnlyHint=false, openWorldHint=true, idempotentHint=false), the description exposes critical behavioral traits: it spends paid quota on every call, runs in a throwaway worktree with a non-applied diff, commits to the spend even if never polled, blocks network egress for sandbox commands, and sends the raw task and read file contents to OpenAI. It also discloses skill auto-loading and best-effort secret redaction, which is significant context not captured in the annotations.

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?

Though lengthy, the description is tightly organized with bolded section cues (PAID, NO NETWORK) and every sentence carries essential operational, cost, privacy, or lifecycle information. The core purpose is front-loaded in the first sentence, and the structure makes the long form scannable. Given the high-stakes paid async nature and the need to disclose privacy limitations, the length is justified and not wasteful.

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?

With an output schema present, the description does not need to detail the return values; it covers everything else needed for correct invocation: cost, async behavior, prerequisites, error-prone network constraints, privacy boundaries, deadlines, and the post-job polling/consumption/cancellation flow. The tool's complexity is high, and this description fully equips an agent to use it safely and effectively.

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?

The input schema already provides 100% coverage with detailed descriptions for all six parameters, so the baseline is 3. The description goes beyond the schema by reinforcing that `workspace_root` must be absolute and, more importantly, warns that the `task` is sent raw to OpenAI and that secret redaction does not cover it. This adds security-relevant meaning to the task parameter that the schema does not express.

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 opens with a precise statement: 'Delegate a coding task to Codex in the background and get a `job_id` back immediately (does not block on the run).' This clearly identifies the action, the resource, and the async behavior, and it distinguishes this tool from the synchronous `codex_delegate` sibling. The paid/quota warning and the contrast with `codex_delegate_dry_run` further disambiguate its purpose.

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 gives explicit when-to-use guidance: prefer this for substantial multi-file tasks that may exceed the 300s synchronous deadline, and use free dry-run/status checks first for scope/readiness. It also names the follow-up lifecycle tools (`codex_job_status`, `codex_job_result`, `codex_job_cancel`) and states the git-repo prerequisite and `workspace_root` requirement, making tool selection and invocation unambiguous.

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