CodeAFTour

Build, audit and review with CodeAF

Built-in specialists: /senior-dev for coding, /sec for security audits and /review for GitHub pull requests.

InstallOne line, then run codeaf in any project.curl -fsSL https://agentfield.ai/get/codeaf | bash
Make your first teamA team with a manager, in five minutes.
Try a playbookCopy a setup we ran, with its time and cost.

The ideas

A few seconds of the product each. Open one to see how it works.

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings 2 want you

teams ▾ dunning emails partial refunds rate limiting × +17 + All

────────────────────────────────────────────────────────────────────────────────────────────────

· new conversation · ~/src/shop-api · crew frugal │ ❯ ctrl+g hide

· 19 other chats are open in this window · alt+v shows them │ rate limit /cart/apply #1

all │ ├─ read the cart routes #2

│ ├─ token bucket in app/… #3

› add rate limiting to /cart/apply and test it │ │ ├─ 10 a minute per AP… #4

│ │ └─ 429 with Retry-Aft… #5

▸ planned the work (7 subtasks) │ └─ tests: burst, refill… #6

│ └─○ tests/test_rate_li… #7

Planned it as seven subtasks. The limiter is a token bucket in │ + /task

app/limits.py, 10 requests a minute per API key, and │

/cart/apply answers 429 with Retry-After when it runs dry. The │ + /standing

tests go in as each piece lands. │

│

│ 2 waiting · 2 done

│ ctrl+. earlier

─ glm-5.3-flash:auto · ◇ YOLO · crew frugal ──── $0.04 · 64% cached idle project: ~/…/shop-api ─

›

opt+e effort · opt+a approvals · opt+k chats · / commands · [ ] tabs All ■■■■■■■■■■■■■■■■■■■□

The Wall

Every conversation, live, at once.

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings 2 want you

teams ▾ dunning emails partial refunds rate limiting +17 + All

Conversations · open in this window 12 running ? 2 needs you 20 open

All 20 │ ● storefront 7 │ ● lexer parser 4 │ › ● evaluator 2 │ ● billing 6

+ New team ✦ Organize

┌─ ◆ storefront manager ─────┐ ┌─ ● fix failing cart t… ──┐ ┌─ ● changelog entry for … ──┐

│ watching 6 members │ │ done · 4m ago │ │ running bash · 2m 00s │

│ @checkout is on th… ▅▆█▄▅▆ │ │ pytest -q · 19 pas… ▃▁▃▂▁▄ │ │ writing CHANGELOG.… ▆▃▄█▄▃ │

└─ Open ↵ ── Select ␣ ───────┘ └────────────────────────────┘ └────────────────────────────┘

┌─ ● ? pricing: round dis… ──┐ ┌─ ● docs/API.md from car… ──┐ ┌─ ● property tests for… ──┐

│ ? waiting on you │ │ thinking · 22s │ │ done · 12m ago │

│ needs you Answer ↵ ▄▁▄▁▃▁ │ │ reading app/cart.py ▂█▄▇▃▆ │ │ 25 passed, three r… ▂▄▄▃▄▂ │

└────────────────────────────┘ └────────────────────────────┘ └────────────────────────────┘

┌─ ● review cart edge cas… ──┐ ┌─ ◆ lexer parser manager ───┐ ┌─ ● fix off by one in … ──┐

│ updated 3m ago │ │ thinking · 9s │ │ done · 6m ago │

│ 11 findings, 2 fix… ▂▂▂▁▃▂ │ │ waiting on @bench ▇▃██▂▄ │ │ go test ./... ok ▄▄▂▂▂▄ │

› / filters · enter opens · m teams

esc back · / filter · m teams · n next that needs you All ■■■■■■■■■■■■■■■■■■■■

Teams

Group conversations into named teams.

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings thu 2:10pm

────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────

● All teams │╭─ Global manager ───────────────────────────────────────────────────────────────────────────────────────────────╮

● storefront ◆ ││ Optional · coordinates the managers of top-level teams │

● api ◆ ││ + Global manager │

● billing ◆ │╰────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯

│╭──────────────────────────────────────────────────────╮ ╭──────────────────────────────────────────────────────╮

││ ● storefront │ │ ● billing │

││ 3 conversations │ │ 2 conversations │

││ 1 running │ │ 1 waiting on you │

││ │ │ │

││ Manager @lead · open model │ │ Manager @billing · open model │

││ Tests and API docs for the rate limit │ │ Choose the invoice rounding rule │

││ Tests passed. API examples are in progress. │ │ Round totals half up or to even? │

││ │ ╰──────────────────────────────────────────────────────╯

││ ▾ Subteams of storefront · 1 │

││ │ ╭────────────────────────────────────────────────╮ │

││ └─│ ● api │ │

││ │ 2 conversations │ │

││ │ 1 running │ │

││ │ Manager @api │ │

││ ╰────────────────────────────────────────────────╯ │

│╰──────────────────────────────────────────────────────╯

│

▫ Show closed · 1 │

Every team, and what waits on you from any of them · enter · tab next place

Managers

Managers, members and recent interactions.

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings 2 want you

teams ▾ dunning emails partial refunds rate limiting × +17 + All

· new conversation · ~/src/shop-api · crew frugal │ ❯ ctrl+g hide

· 19 other chats are open in this window · alt+v shows them │ rate limit /cart/apply #1

all │ ├─ read the cart routes #2

│ ├─ token bucket in app/… #3

› add rate limiting to /cart/apply and test it │ │ ├─ 10 a minute per AP… #4

│ │ └─ 429 with Retry-Aft… #5

▸ planned the work (7 subtasks) │ └─ tests: burst, refill… #6

│ └─○ tests/test_rate_li… #7

Planned it as seven subtasks. The limiter is a token bucket in │ + /task

app/limits.py, 10 requests a minute per API key, and │

/cart/apply answers 429 with Retry-After when it runs dry. The │ + /standing

tests go in as each piece lands. │

│

│ 2 waiting · 2 done

│ ctrl+. earlier

─ glm-5.3-flash:auto · ◇ YOLO · crew frugal ──── $0.04 · 64% cached idle project: ~/…/shop-api ─

› /senior-dev make checkout idempotent across retries

opt+e effort · opt+a approvals · opt+k chats · / commands All ■■■■■■■■■■■■■■■■■■■□

Specialists

Built-in specialists: build, audit and review.

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings 2 want you

teams ▾ dunning emails partial refunds rate limiting × +17 + All

· new conversation · ~/src/shop-api · crew frugal │ ❯ ctrl+g hide

· 19 other chats are open in this window · alt+v shows them │ rate limit /cart/apply #1

all │ ├─ read the cart routes #2

│ ├─ token bucket in app/… #3

› add rate limiting to /cart/apply and test it │ │ ├─ 10 a minute per AP… #4

│ │ └─ 429 with Retry-Aft… #5

▸ planned the work (7 subtasks) │ └─ tests: burst, refill… #6

│ └─○ tests/test_rate_li… #7

Planned it as seven subtasks. The limiter is a token bucket in │ + /task

app/limits.py, 10 requests a minute per API key, and │

/cart/apply answers 429 with Retry-After when it runs dry. The │ + /standing

tests go in as each piece lands. │

│

│ 2 waiting · 2 done

│ ctrl+. earlier

─ glm-5.3-flash:auto · ◇ YOLO · crew frugal ──── $0.04 · 64% cached idle project: ~/…/shop-api ─

›

opt+e effort · opt+a approvals · opt+k chats · / commands · [ ] tabs All ■■■■■■■■■■■■■■■■■■■□

Pareto crewing

Every seat gets its model.

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings 2 want you

teams ▾ dunning emails partial refunds rate limiting × +17 + All

· new conversation · ~/src/shop-api · crew frugal │ ❯ ctrl+g hide

· 19 other chats are open in this window · alt+v shows them │ rate limit /cart/apply #1

all │ ├─ read the cart routes #2

│ ├─ token bucket in app/… #3

› add rate limiting to /cart/apply and test it │ │ ├─ 10 a minute per AP… #4

│ │ └─ 429 with Retry-Aft… #5

▸ planned the work (7 subtasks) │ └─ tests: burst, refill… #6

│ └─○ tests/test_rate_li… #7

Planned it as seven subtasks. The limiter is a token bucket in │ + /task

app/limits.py, 10 requests a minute per API key, and │

/cart/apply answers 429 with Retry-After when it runs dry. The │ + /standing

tests go in as each piece lands. │

│

│ 2 waiting · 2 done

│ ctrl+. earlier

─ glm-5.3-flash:auto · ◇ YOLO · crew frugal ──── $0.04 · 64% cached idle project: ~/…/shop-api ─

› every Monday at 9, draft the weekly update

opt+e effort · opt+a approvals · opt+k chats · / commands All ■■■■■■■■■■■■■■■■■■■□

Standing orders

Say it once; it keeps running.

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings 2 want you

teams ▾ dunning emails partial refunds link expiry × +17 + All

· new conversation · shop-api │ ❯ ctrl+g hide

│ Add link expiry and c… 33s

· single task 1 started · Add link expiry and click stats to │ + /task

─ glm-5.3-flash:auto · ◇ YOLO ───────────── $0.04 · 64% cached working project: ~/…/shop-api ─

›

ctrl+c interrupt · opt+k chats · / commands · [ ] tabs All ■■■■■■■■■■■■■■■■■■■□

Keeps working

Close the window; work continues.

# a CI step, nobody at the keyboard

$ codeaf do "Make the failing tests pass by fixing the code, not the tests." --json

Headless mode

Hand it a job from anywhere.

The playbook wall

Setups we ran end to end, with the time and the bill of our run. Copy one today.

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings 2 want you

teams ▾ partial refunds Give all 100 produc... × +18 + All

· new conversation · shop-api │ ❯ ctrl+g hide

─ glm-5.3-flash:auto · ◇ YOLO ────────────────── $0.04 · 64% cached idle project: ~/…/shop-api ─

›

opt+e effort · opt+a approvals · opt+k chats · / commands · [ ] tabs All ■■■■■■■■■■■■■■■■■■■□

Go wide

Run a hundred jobs at once

One brief becomes a hundred parts, one per file, and all hundred run side by side under a cap you set.

2 min 26 s · $1.52 in our run · Intermediate

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings 2 want you

teams ▾ dunning emails partial refunds link expiry × +17 + All

· new conversation · shop-api │ ❯ ctrl+g hide

│ Add link expiry and cl… 1m

· done · ran 1m 6s · Added TTL expiry and click stats … │ + /task

Evidence: python3 -m unittest -> Ran 9 tests, OK. │

◆ Add link expiry and click stats to linkshort · done · 1m06s │ + /standing

· 9 files · branch kept │

│

│ 0 waiting · 1 done

│ ctrl+. earlier

─ glm-5.3-flash:auto · ◇ YOLO ────────────────── $0.04 · 64% cached idle project: ~/…/shop-api ─

›

opt+e effort · opt+a approvals · opt+k chats · / commands · [ ] tabs All ■■■■■■■■■■■■■■■■■■■□

Run while you sleep

Ship a feature while you are away

Send one brief, close the terminal, and come back to a checked branch in the same conversation.

1 min 6 s · $0.16 in our run · Starter

$ claude -p "Add an aging report … Follow CLAUDE.md."

claude: codeaf do "Add an 'aging' subcommand … Done means: python3 -m unittest passes." --json

From anywhere

Claude Code plans, CodeAF builds

Claude Code plans and reviews, and CodeAF's crew writes the implementation.

3 min 34 s · $0.20 in our run · Intermediate

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings 2 want you

● storefront ▾ ◆ Manager × property tests cart edge cases +4 + All

· ◆ storefront manager · asks you only what it cannot decide │ Tasks 0 · Traffic 4 alt+l

│

› keep the cart and checkout green while the pricing work │ ◆ → @cart done

lands │ fix the failing cart tests

│

▸ managed the team (team_status, team_read @cart @checkout) │ ◆ → @checkout running ▸

│ changelog entry for the c…

Team status: @cart fixed the failing cart tests, 19 passed. │ └ @checkout → ◆ drafting …

@checkout is writing the changelog entry for that fix. @docs │

is drafting docs/API.md from cart.py. │ ? @pricing → you asking

│ round discounts half up o…

One thing is waiting on you: @pricing asks whether discounts │

round half up or to even. │ ◆ → @docs running ▸

│ write docs/API.md from ca…

│

─ glm-5.3-flash:auto · ◇ YOLO ────────────────── $0.04 · 64% cached idle project: ~/…/shop-api ─

› to ◆ storefront manager

opt+e effort · opt+a approvals · opt+k chats · / commands · [ ] tabs All □■■■■■■■■■■■■■■■■■■■

Beyond code

Make a launch campaign from one brief

A small team turns a product brief into copy, a banner, a social picture and a voiceover in a few minutes.

3 min · $0.34 in our run · Intermediate

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings 2 want you

teams ▾ dunning emails partial refunds rate limiting × +17 + All

· new conversation · ~/src/shop-api · crew frugal │ ❯ ctrl+g hide

· 19 other chats are open in this window · alt+v shows them │ rate limit /cart/apply #1

all │ ├─ read the cart routes #2

│ ├─ token bucket in app/… #3

› add rate limiting to /cart/apply and test it │ │ ├─ 10 a minute per AP… #4

│ │ └─ 429 with Retry-Aft… #5

▸ planned the work (7 subtasks) │ └─ tests: burst, refill… #6

│ └─○ tests/test_rate_li… #7

Planned it as seven subtasks. The limiter is a token bucket in │ + /task

app/limits.py, 10 requests a minute per API key, and │

/cart/apply answers 429 with Retry-After when it runs dry. The │ + /standing

tests go in as each piece lands. │

│

│ 2 waiting · 2 done

│ ctrl+. earlier

─ glm-5.3-flash:auto · ◇ YOLO · crew frugal ──── $0.04 · 64% cached idle project: ~/…/shop-api ─

›

opt+e effort · opt+a approvals · opt+k chats · / commands · [ ] tabs All ■■■■■■■■■■■■■■■■■■■□

Run while you sleep

Run an on-call pager

A watch checks your service every few minutes, and when it fails, a worker finds the cause, fixes it and logs the incident.

45 s · $0.03 in our run · Starter

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings 2 want you

teams ▾ dunning emails partial refunds tempo v1.3 trailer × +17 + All

· new conversation · tempo │ ❯ ctrl+g hide

│ + /task

› Make a release trailer for tempo v1.3 from git log │

v1.2..HEAD … │ + /standing

Loaded: generate_image, speak, generate_music, generate_video, │

edit_video │

media/voice.mp3, 177.1KB of mp3 audio, spoken by │

fish-audio/s2.1-pro │

job 1 started; composing on google/lyria-3-clip-preview │

media/title.png, 2048×1152, generated on │

bytedance-seed/seedream-5-0-pro │

job 2 started; filming on bytedance/seedance-2.0-mini │

│

│ ctrl+. earlier

─ glm-5.3-flash:auto · ◇ YOLO ─ $0.04 · saved $0.0100 · 64% cached idle project: ~/src/tempo ─

›

opt+e effort · opt+a approvals · opt+k chats · / commands · [ ] tabs All ■■■■■■■■■■■■■■■■■■■□

Beyond code

Cut a release trailer from the changelog

One brief turns git log into release notes, a voiceover, music, a title card and a clip, cut into one trailer.

4 min 32 s · $0.44 in our run · Intermediate

>● codeaf Home Chats AI teams Activity Memories 1 Automations Spend Settings 2 want you

teams ▾ dunning emails partial refunds rate limiting +17 + All

Conversations · open in this window 12 running ? 2 needs you 20 open

All 20 │ ● storefront 7 │ ● lexer parser 4 │ › ● evaluator 2 │ ● billing 6

+ New team ✦ Organize

┌─ ◆ storefront manager ─────┐ ┌─ ● fix failing cart t… ──┐ ┌─ ● changelog entry for … ──┐

│ watching 6 members │ │ done · 4m ago │ │ running bash · 2m 00s │

│ @checkout is on th… ▅▆█▄▅▆ │ │ pytest -q · 19 pas… ▃▁▃▂▁▄ │ │ writing CHANGELOG.… ▆▃▄█▄▃ │

┌─ ● ? pricing: round dis… ──┐ ┌─ ● docs/API.md from car… ──┐ ┌─ ● property tests for… ──┐

│ ? waiting on you │ │ thinking · 22s │ │ done · 12m ago │

│ needs you Answer ↵ ▄▁▄▁▃▁ │ │ reading app/cart.py ▂█▄▇▃▆ │ │ 25 passed, three r… ▂▄▄▃▄▂ │

┌─ ● review cart edge cas… ──┐ ┌─ ◆ lexer parser manager ───┐ ┌─ ● fix off by one in … ──┐

│ updated 3m ago │ │ thinking · 9s │ │ done · 6m ago │

│ 11 findings, 2 fix… ▂▂▂▁▃▂ │ │ waiting on @bench ▇▃██▂▄ │ │ go test ./... ok ▄▄▂▂▂▄ │

› / filters · enter opens · m teams

esc back · / filter · m teams · n next that needs you All ■■■■■■■■■■■■■■■■■■■■

Build an org

Build a software company

Write one message to a global manager and a company of teams plans, builds and tests a feature.

12 min · $0.70 in our run · Advanced

# textkit: a teammate's commit broke 2 of 5 tests

$ git push origin main

From anywhere

Fix the build before you push

A git pre-push hook runs the tests, and when they fail, codeaf do fixes the code and the next push carries the fix.

14 s · $0.006 in our run · Intermediate

All playbooks