[SYMBaiEX]

Summary

• Extract Agent Fighter local policy decisions into a pure module and add isolated, per-agent mulberry32-v1 streams for seeded local choices without modifying global Math.random.
• Add --seed <uint32> for runner-spawned servers and expose seed/RNG scope metadata in server telemetry and headless JSON output. Reject --seed with --url; external server seeds are not reported by the runner.
• Add seeded-policy tests to the root test workflow and document the observation, provider, timing, runtime, and version limits.

The seed controls only randomized local-policy branches. OpenAI Responses API outcomes and whole-match behavior are not deterministic claims in this change.

Validation

• bun test examples/agent-fighter/__tests__/seeded-policy.test.ts — 4 passed.
• bun run release:check — passed; 166 tests, 0 failures, typecheck/build/audit/package checks passed, no audit vulnerabilities.
• bun --cwd examples/agent-fighter headless --seed 42 --duration-ms 1000 --poll-ms 100 --min-decisions 0 --min-decisions-per-player 0 --port 53143 --output /tmp/opencontroller-agent-fighter-seed-check.json — passed; summary and server telemetry both reported seed 42 and mulberry32-v1; 16 decisions used the local policy. The smoke used the existing system Chromium because the Playwright CDN download was blocked in this environment.
• git diff --check — passed.

Closes #22

Summary by CodeRabbit

• New Features
  • Added optional seed support for repeatable local fighter-policy decisions in headless matches and runner-started servers.
  • Headless summaries now report the seed and random-source details when available. External-server runs identify the random source as unknown when the runner cannot verify it.
  • The --seed option cannot be used with --url.
• Documentation
  • Added guidance on seed setup and clarified that seeded decisions do not guarantee identical full-match results.

[Copilot]

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

[coderabbitai]

Navigate logical layers of code changes, visualize relationships, and explore their blast radius.

Note

Currently processing new changes in this PR. This may take a few minutes, please wait...

⚙️ Run configuration

• Configuration used: defaults
• Review profile: CHILL
• Plan: Advanced
• Run ID: b5b8b295-c649-41b1-b5d5-40c872debe33

📥 Commits

Reviewing files that changed from the base of the PR and between 4cef8d7 and 7efced7.

📒 Files selected for processing (9)

• docs/agent-fighter.md
• examples/agent-fighter/README.md
• examples/agent-fighter/__tests__/seeded-policy.test.ts
• examples/agent-fighter/headless-options.ts
• examples/agent-fighter/headless.ts
• examples/agent-fighter/local-policy.ts
• examples/agent-fighter/server.ts
• examples/agent-fighter/tsconfig.json
• package.json

 _______________________________________________________________________________
< Fancy algorithms are slow when n is small, and n is usually small. - Rob Pike >
 -------------------------------------------------------------------------------
  \
   \   (\__/)
       (•ㅅ•)
       / 　 づ

✨ Finishing Touches

📝 Generate docstrings

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

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.}