fix(conformance): bind Chat launch readiness budget

[BunsDev]

Summary

• bind merged fix(conformance): align native launch RPC timeout chat#243 and its updated protected workflow digest
• refresh the exact Windows bootstrap, nested child bootstrap, and Unix supervisor preparation fixtures
• preserve fail-closed workflow verification while documenting the command-specific launch budget

Validation

• corepack pnpm@10.34.0 --ignore-workspace run verify

[Copilot]

🟡 Changes recommended

Unresolved timer and error-mapping issues affect the bound producer, with related documentation requiring correction.

Get a fresh assessment by requesting another Copilot review.

Pull request overview

Rebinds cross-repository conformance evidence to merged Chat #243, refreshing integrity fixtures and documenting the cave_launch budget.

Changes:

• Updates producer, workflow, bootstrap, and supervisor integrity data.
• Refreshes conformance tests and lock metadata.
• Documents fail-closed verification and launch-budget expectations.

Open findings:

• RELEASING.md — nit (1 vote): the harness still uses a fixed 10-second timer.
• conformance/client-v1-cross-repository-lock.json — moderate (3 votes): a second client still uses the 10-second timer.
• conformance/client-v1-cross-repository-lock.json — moderate (1 vote): all service_unavailable failures are mapped to launch timeouts.
• docs/workflows/client-v1-cross-repository-conformance.md — nit (1 vote): the documented mapping is too broad.

File summaries

File	Summary
tests/conformance-gaps.spec.ts	Updates frozen producer and fixture expectations.
tests/conformance-contract.spec.ts	Refreshes producer identities and contract assertions.
scripts/github-conformance-evidence.mjs	Updates reviewed integrity pins.
RELEASING.md	Documents the producer binding and launch budget.
docs/workflows/client-v1-cross-repository-conformance.md	Updates conformance binding documentation.
conformance/client-v1-cross-repository-lock.json	Rebinds the frozen producer evidence metadata.

Review details

Suppressed comments (3)

RELEASING.md:222

• This release note overstates the rebind: the frozen test:phase1-conformance harness still uses a fixed 10-second RPC timer for its cave_launch call, while Chat #243 only gives the separate schema-v2 producer client a 35-second budget. Update the harness and rebind its integrity data before documenting this as covering the native readiness deadline.

`34717180149` passed its applicable checks. The producer gives only
`cave_launch` a 35-second RPC response budget around the native 30-second
readiness deadline; all other RPCs retain their 10-second bound. The merge

conformance/client-v1-cross-repository-lock.json:168

• The bound producer also maps every native RPC cave_launch failed with service_unavailable to phase1.native-scenarios.launch.timeout, but Chat #243's independent review notes that service_unavailable is also emitted for worker-completion loss, blocking-task completion loss, and discovery-generation overflow. This can mislabel non-timeout failures as readiness expiry in the evidence; the producer should require explicit timeout provenance before this commit is marked compatible.

    "commit": "59fd228e308b732e5f530b0f89af68f3811a177c",

docs/workflows/client-v1-cross-repository-conformance.md:36

• The documented mapping is narrower than the bound producer's behavior. Chat #243 maps every native RPC cave_launch failed with service_unavailable to phase1.native-scenarios.launch.timeout, but service_unavailable also represents worker-completion loss, blocking-task completion loss, and discovery-generation overflow. Those failures will be mislabeled as readiness timeouts; constrain the mapping to explicit readiness-expiry provenance (or retain a separate service-failure diagnostic) before locking this producer.

35-second outer RPC response budget needed to encompass the native 30-second
readiness deadline; every other RPC retains its 10-second bound. The merge

• Files reviewed: 6/9 changed files
• Comments generated: 1
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.