[codex] Explore Codex CLI app-server feasibility

[hannasdev]

Summary

This PR publishes the Codex CLI/app-server feasibility spike as an experimental product direction, not a polished replacement workflow.

It adds repeatable probes and documentation for three usage paths:

• Claude Code wrapper: the most complete current Switchboard path.
• Advisory cross-surface routing: recommendation-only routing for other clients.
• Codex app-server spike: experimental Switchboard-owned session surface for in-session route/model override exploration.

What changed

• Added Codex CLI command-surface and live resume probes.
• Added Codex app-server switch, protocol, preflight, lifecycle, and model-evidence probes.
• Added focused tests for protocol shape, preflight failure modes, lifecycle safety, and app-server switch behavior.
• Updated the spike scope and decision log through all supportability gates.
• Reframed the README for early adopters, making clear what is usable today and what remains experimental.

Gate status

• Gate 1 Public Surface: pass for bounded experimental spike.
• Gate 2 Protocol Stability: pass with generated-binding compatibility check.
• Gate 3 User Install And Auth Path: pass with preflight diagnostics.
• Gate 4 Process Lifecycle Safety: pass with live lifecycle probe plus fake-process failure coverage.
• Gate 5 Continuity And Session Semantics: pass for same app-server thread/session.
• Gate 6 Model Evidence: accepted risk. We prove requested model override plus same-thread completion, but not provider-side backend model attestation.
• Gate 7 Product Fit: pass for experimental publication. Product accepts a Switchboard-owned Codex app-server session surface and does not claim stock Codex TUI hot-swapping.

Validation

• npm run lint
• npm test
• npm run switchboard:spike:codex-app-server:preflight
• npm run switchboard:spike:codex-app-server:lifecycle
• npm run switchboard:spike:codex-app-server

Live Codex app-server commands require normal local Codex auth/state and were run outside the sandbox where needed.

Remaining caveats

• Codex app-server is still experimental.
• There is no polished Switchboard-owned Codex UI yet.
• This does not prove hot-swapping inside the stock Codex TUI.
• Backend/effective model telemetry remains unresolved; current claims must stay framed as experimental feasibility evidence.

[Copilot]

Pull request overview

This PR adds a bounded Codex CLI/app-server feasibility spike alongside the existing Claude-focused Switchboard workflow, documenting what is usable today versus experimental.

Changes:

• Adds Codex CLI and app-server probe scripts for command surface, protocol compatibility, preflight/auth, lifecycle, and in-session model override evidence.
• Adds automated tests with fake Codex binaries/app-server processes for the new probes.
• Updates README, product scope, PRD index, decision log, package scripts, and agent onboarding docs to publish the spike framing.

Reviewed changes

Copilot reviewed 16 out of 16 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
test/codex-cli-feasibility-probe.test.js	Tests CLI command-surface and live resume probe behavior.
test/codex-app-server-switch-probe.test.js	Tests app-server one-thread model override probe.
test/codex-app-server-protocol-check.test.js	Tests protocol shape validation and failure reporting.
test/codex-app-server-preflight.test.js	Tests Codex version, app-server availability, and auth preflight paths.
test/codex-app-server-lifecycle-probe.test.js	Tests lifecycle, crash, malformed output, and spawn-failure handling.
scripts/codex-cli-feasibility-probe.js	Adds Codex CLI route/resume feasibility probe.
scripts/codex-app-server-switch-probe.js	Adds app-server model override feasibility probe.
scripts/codex-app-server-protocol-check.js	Adds generated-binding protocol compatibility check.
scripts/codex-app-server-preflight.js	Adds install/auth readiness diagnostics for app-server usage.
scripts/codex-app-server-lifecycle-probe.js	Adds process lifecycle safety probe.
README.md	Reframes product usage paths and documents spike commands.
package.json	Adds npm scripts for the new spike probes.
docs/product/CODEX-CLI-SPIKE-SCOPE.md	Adds detailed spike scope, gates, evidence, and decision criteria.
docs/PRD.md	Links the Codex spike scope from the PRD index.
docs/decision-log.md	Records the Codex feasibility spike decision and evidence.
AGENTS.md	Adds the Codex spike scope to recommended reading order.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 16 out of 16 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 16 out of 16 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

Copilot reviewed 16 out of 16 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 16 out of 16 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

Copilot reviewed 16 out of 16 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

Copilot reviewed 16 out of 16 changed files in this pull request and generated 3 comments.

[Copilot]

Pull request overview

Copilot reviewed 16 out of 16 changed files in this pull request and generated 1 comment.

[Copilot]

Pull request overview

Copilot reviewed 16 out of 16 changed files in this pull request and generated no new comments.