Skip to content

feat: Add daily performance agent - #13729

Merged
anthonyshew merged 5 commits into
mainfrom
daily-performance-agent
Aug 12, 2026
Merged

anthonyshew merged 5 commits into
mainfrom
daily-performance-agent

Conversation

@anthonyshew

@anthonyshew anthonyshew commented Aug 12, 2026

Copy link
Copy Markdown
Contributor

Summary

  • Add a daily Eve workflow for focused Turborepo performance improvements.
  • Alternate GPT 5.6 Sol and Claude Fable 5 authorship by UTC day, with opposite-model adversarial reviewers.
  • Require optimized baseline and candidate binaries, balanced AB/BA measurements, controlled cache state, and a 95% confidence interval above zero; every reproducible millisecond counts.
  • Fingerprint benchmark binaries and immutable workload revisions, and invalidate stale comparison evidence when the diff changes.
  • Document turbo --profile, DHAT heap analysis, Linux strace, and pinned open-source Turborepos spanning small through very large repositories, including pnpm, Yarn, and Bun.
  • Enforce correctness validation, adversarial review, protected-path, and draft-PR gates.
  • Preserve the existing daily examples-maintenance schedule.

Verification

  • pnpm --dir apps/agents test
  • pnpm --dir apps/agents exec tsc --noEmit
  • pnpm --dir apps/agents exec eve info --json
  • pnpm --dir apps/agents exec eve build
  • pnpm --dir apps/agents build
  • Repository pre-push format and TOML checks

🤖 Generated with Claude Code

Co-Authored-By: Claude <noreply@anthropic.com>

Co-Authored-By: Anthony Shew <35677084+anthonyshew@users.noreply.github.com>
@anthonyshew
anthonyshew requested review from a team and tknickman August 12, 2026 13:59
@vercel

vercel Bot commented Aug 12, 2026

Copy link
Copy Markdown
Contributor

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated (UTC)
examples-basic-web Ready Ready Preview Aug 12, 2026 5:57pm
examples-designsystem-docs Ready Ready Preview Aug 12, 2026 5:57pm
examples-gatsby-web Ready Ready Preview Aug 12, 2026 5:57pm
examples-kitchensink-blog Ready Ready Preview Aug 12, 2026 5:57pm
examples-nonmonorepo Ready Ready Preview Aug 12, 2026 5:57pm
examples-svelte-web Ready Ready Preview Aug 12, 2026 5:57pm
examples-tailwind-web Ready Ready Preview Aug 12, 2026 5:57pm
examples-vite-web Ready Ready Preview Aug 12, 2026 5:57pm
turbo-site Ready Ready Preview Aug 12, 2026 5:57pm
turborepo-eve-agent Ready Ready Preview Aug 12, 2026 5:57pm

Comment thread apps/agents/agent/lib/performance-validation.ts
## Summary

Stacked on #13729 — base is `daily-performance-agent`, so this diff is
only the operator surface for that workflow.

- add a **Daily performance improvement** card to the Eve operator
dashboard with a `Run performance improvement now` control, the 15:30
UTC schedule, the rotation rule, and the review/output gates
- dispatch operator runs by `x-operator-action`: the new performance
action sends the same `DAILY_PERFORMANCE_IMPROVEMENT_PROMPT` the
schedule uses, with app auth and task mode, so
`begin_performance_improvement`, the opposite-model adversarial review,
the stale-diff/protected-path checks, and draft-PR publication all apply
to manual runs
- report the author and reviewer models for a started run, resolved with
`selectPerformanceModels(sessionDate(session.id))` so the dashboard
shows the same pair the agent resolves rather than a separate guess
- share the operator action names in `agent/lib/operator-runs.ts` so the
dashboard and channel cannot drift, and extract the start/poll/status
logic into `app/operator-run.tsx` so both operation cards behave
identically
- extend the standing instructions to cover operator performance runs,
matching how example maintenance already reads

Existing behavior is unchanged: the maintenance action keeps its origin,
action, content-type, and example validation, and its start response is
byte-identical (the optional `models` field is omitted).

## Verification

- `pnpm --dir apps/agents test`
- `pnpm --dir apps/agents exec tsc --noEmit`
- `pnpm --dir apps/agents build`
- `pnpm --dir apps/agents exec eve info` reports 0 diagnostics and
discovers both operator routes, both schedules, and both reviewer
subagents
- `next dev` render check: both operation cards, facts, and controls
render; the client bundle carries both operator actions
- against the running channel: missing origin, unknown action, and
non-JSON content type each return `403 Invalid operator request.`; a
matching origin with an invalid example returns `400 A valid example is
required.`
- repository pre-push hooks (lint-staged, `turbo run format check:toml`,
`cargo fmt --check`) and `oxlint apps/agents`

A live performance run was deliberately not triggered from this box: it
would clone the repo, spend model budget, and can publish a draft pull
request.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: vercel[bot] <35613825+vercel[bot]@users.noreply.github.com>
@anthonyshew
anthonyshew merged commit c4ac03b into main Aug 12, 2026
78 of 82 checks passed
@anthonyshew
anthonyshew deleted the daily-performance-agent branch August 12, 2026 22:19
anthonyshew pushed a commit that referenced this pull request Aug 14, 2026
## Release v2.10.10

> [!CAUTION]
> Versioned docs aliasing FAILED. [View
logs](https://github.com/vercel/turborepo/actions/runs/31803665773)

### Changes

- chore: Release Turborepo 2.10.9 (#13696) (`c21e2e2`)
- docs: Update Geistdocs to 1.19.6 (#13701) (`9bbc2c8`)
- feat: Support nub.lock files (#13699) (`4740bac`)
- chore: Release Turborepo 2.10.10-canary.1 (#13703) (`44c6fd6`)
- perf: Walk the repository once when pruning (#13705) (`c59a533`)
- fix(deps): Upgrade js-yaml to 4.3.1 (GHSA-5p4m-2wfm-xmqj) (#13704)
(`f12c6e9`)
- chore: Move pnpm overrides to pnpm-workspace.yaml (#13706) (`1475441`)
- chore: Release Turborepo 2.10.10-canary.2 (#13707) (`c72285f`)
- fix: Enable Eve examples pull requests (#13708) (`4e57b15`)
- feat: Add Eve operator dashboard (#13710) (`96b9647`)
- fix: Remove unsupported sandbox network policy (#13712) (`688614d`)
- feat: Rotate Eve example maintenance daily (#13715) (`ec56a7f`)
- fix: Resolve agent lint errors (#13716) (`bb535e9`)
- chore: Upgrade TypeScript to 7.0.2 (#13713) (`2afa75d`)
- fix(cli): Move EXPERIMENTAL label from ls command to --output flag
(#13709) (`422ba28`)
- fix: Apply nested parent gitignore patterns in manual hashing (#13690)
(`2b8863c`)
- feat: Add pytest task discovery (#13720) (`24e07a7`)
- chore: Release Turborepo 2.10.10-canary.3 (#13721) (`bb46749`)
- chore: Validate updated examples with turbo (#13723) (`ad53149`)
- chore: Update with-rollup example (#13718) (`42ec389`)
- feat: Notify Slack when agents open PRs (#13722) (`aa070b1`)
- chore: Enforce draft pull requests (#13726) (`6a46551`)
- fix: Expose Slack delivery diagnostics (#13733) (`caddaa9`)
- chore: Update non-monorepo example (#13727) (`0bfd752`)
- chore: Update with-rsbuild example (#13730) (`e0aee94`)
- feat: Add daily performance agent (#13729) (`c4ac03b`)
- feat: Support cross-toolchain repository affectedness (#13737)
(`31b0c4f`)
- fix: Scope musl library dependency installation (#13738) (`1da105f`)
- chore: Release Turbo repository packages 0.0.1-canary.24 (#13739)
(`65aa565`)
- fix(bun): Preserve overrides objects, trustedDependencies, workspace
bins and git integrity through prune; accept lockfileVersion 3 (#13740)
(`ea08fac`)

---------

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
github-actions Bot added a commit that referenced this pull request Aug 14, 2026
## Release v2.10.11-canary.1

> [!CAUTION]
> Versioned docs aliasing FAILED. [View
logs](https://github.com/vercel/turborepo/actions/runs/31812668548)

### Changes

- chore: Release Turborepo 2.10.10-canary.3 (#13721) (`bb46749`)
- chore: Validate updated examples with turbo (#13723) (`ad53149`)
- chore: Update with-rollup example (#13718) (`42ec389`)
- feat: Notify Slack when agents open PRs (#13722) (`aa070b1`)
- chore: Enforce draft pull requests (#13726) (`6a46551`)
- fix: Expose Slack delivery diagnostics (#13733) (`caddaa9`)
- chore: Update non-monorepo example (#13727) (`0bfd752`)
- chore: Update with-rsbuild example (#13730) (`e0aee94`)
- feat: Add daily performance agent (#13729) (`c4ac03b`)
- feat: Support cross-toolchain repository affectedness (#13737)
(`31b0c4f`)
- fix: Scope musl library dependency installation (#13738) (`1da105f`)
- chore: Release Turbo repository packages 0.0.1-canary.24 (#13739)
(`65aa565`)
- fix(bun): Preserve overrides objects, trustedDependencies, workspace
bins and git integrity through prune; accept lockfileVersion 3 (#13740)
(`ea08fac`)
- chore: Release Turborepo 2.10.10 (#13741) (`f7cb04d`)
- fix: Automatically approve release workflows (#13743) (`237aef5`)

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant