Use Codex effective context windows

[KeenWill]

Outcome

Codex CLI deployments now budget against the context window the CLI actually admits, rather than its larger raw catalog value. The shipped catalog records the CLI's 95% reservation explicitly: gpt-5.6-sol is 258,400 usable tokens from 272,000 raw, and gpt-5.3-codex-spark is 121,600 from 128,000. The owning configuration contract now says context_window_tokens is the usable ceiling after provider or adapter reservations.

Dogfood also lowered both active gpt-5.6-sol and gpt-5.6-terra declarations from 272,000 to 258,400 before deployment. This closes the 5,408-token blind spot where Signalbox admitted a continuation that Codex rejected as ContextWindowExceeded.

Meaningfully changed lines: 23 (excluding lockfiles).

Bounds

• Corrected capability ceilings: Codex gpt-5.6-sol/terra 272,000 → 258,400 usable tokens; Codex Spark example 128,000 → 121,600.
• No new runtime retry, time, page-size, or concurrency bound.

Validation

• Checked-in example parsing and inactive-catalog activation tests pass.
• Workspace check, strict Clippy, formatting, doctests, rustdoc, metadata, Markdown, numeric-bound, migration, invariant, domain-spine, and checker self-tests pass.
• Full workspace tests retain unrelated host/fixture failures: the IANA alias fixture, local-socket ancestor permission fixtures, and two Codex process timing/ETXTBSY cases; both Codex cases pass in isolation.
• Docs consistency retains the inherited bounds carrier warning for PR Reconcile commissioned approval escalations #1056/Make numeric-bound declarations fail closed #1125.

Live deploy

• Deploy ref and installed stack tip: 1bb1f45a1.
• Clean graceful shutdown; startup completed migration, startup scan, and socket binding. Three fresh active turns are running and delivery receipts advance; event-projection freshness is soaking.
• One pre-existing boundary-loss turn entered bounded model-call reconciliation at startup.

Absorption ledger: none.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: ead82dcb-e20b-481c-a7eb-fc40e90aa69f

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

Warning

Your free Security trial is over. An organization admin can activate billing to continue.

---

_{Comment @coderabbitai help to get the list of available commands.}

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[chatgpt-codex-connector]

You have reached your Codex usage limits for security reviews. Please try again later.