docs(evidence): add #11504 stage1 code-reply proof

[lalalune]

Closes #11504

Summary

The runtime fix for #11504 landed in #11555 (c532e3b565e): isUnusableStage1Reply no longer defers valid bare-code or structured replies just because they contain repeated-character runs. This PR adds the missing live evidence artifact and manual review notes required to close the issue.

Evidence Added

• .github/issue-evidence/11504-stage1-code-replies/README.md
• .github/issue-evidence/11504-stage1-code-replies/live-cerebras-current/93432706-b3b2-08ea-ab6a-ba55340a8848/tj-74b5c8fef7d6a9.json

Manual review of the trajectory confirms:

• status: finished
• single messageHandler stage
• model gemma-4-31b
• selected contexts:["simple"] with requiresTool:false
• returned an unfenced Python code body as direct_reply instead of I'm not sure how to answer that.
• no API key or authorization strings found in the trajectory artifact

Verification

• bun run --cwd packages/core test src/__tests__/message-runtime-stage1.test.ts -> 1 passed, 75 passed
• bunx @biomejs/biome check packages/core/src/services/message.ts packages/core/src/__tests__/message-runtime-stage1.test.ts .github/issue-evidence/11504-stage1-code-replies/README.md -> Checked 2 files ... No fixes applied.
• bun run --cwd packages/core typecheck -> pass
• live gemma-4-31b Cerebras runtime capture -> pass, artifact above
• bun run install:light -> pass, artifact sync skipped with ELIZA_SKIP_ARTIFACT_SYNC=1
• full bun install attempted, but postinstall artifact sync was still at 7.0 MiB / 971 MiB with 141m 59s remaining, so it was stopped before completion
• bun run verify currently fails before typecheck/lint in audit:type-safety-ratchet on current repo baseline (as unknown as: 77 / 76, ?? 0 core/agent/app-core: 376 / 375). This PR changes only evidence artifacts, with no production-source diff.

UI / Media

N/A - core runtime evidence only; no user-facing UI changed.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. 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: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 3bd6ba39-3612-4915-8e07-d0dc27cc00e9

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

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch docs/11504-stage1-evidence

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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

[greptile-apps]

Your trial has ended. Reactivate Greptile to resume code reviews.

[greptile-apps]

Your trial has ended. Reactivate Greptile to resume code reviews.

[lalalune]

Reviewed locally as evidence-only, but GitHub will not let this account approve its own PR. Rebased onto origin/develop and pushed c79bf9c. I validated the trajectory JSON parses and matches the README claims: status=finished, one messageHandler stage, model=gemma-4-31b, contexts=[simple], requiresTool=false, and the replyText is a Python function body rather than the fallback. I also checked the README issue/artifact links, ran git diff --check, and ran a narrow credential-pattern scan for sk-/sk-proj/Bearer/Authorization/x-api-key shaped values with no matches. Biome ignores this .github evidence README in the current config, so there was no meaningful Biome file to check here.

[claude]

Claude encountered an error —— View job

---

I'll analyze this and get back to you.