Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Cognitive Modes · vibeeval | DecimalAI

Cognitive Modes

SkillScore32Maintained

cognitive-modes

Gorev tipine gore dusunme modu secimi. 5 mod -- analytical (derin analiz), creative (yaratici cozum), systematic (adim adim), rapid (hizli aksiyon), debug (hata izleme). Her mod farkli prompt stratejisi ve karar agaci kullanir.

agent-toolingby vibeeval

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

AI Engineer pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/cognitive-modes/SKILL.md

decimalai skills pull cognitive-modes --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/cognitive-modes/SKILL.md https://app.decimal.ai/s/cognitive-modes/SKILL.md

Raw file: /s/cognitive-modes/SKILL.md ↗ · machine-readable summary at /s/cognitive-modes.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 55d ago

Yes — +32% lift · at higher cost+108% tokens0% turns32% of whole cases pass with skill (7/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-16	✗→✓	▲ Improved	—	—
case-19	✗→✓	▲ Improved	—	—
case-12	✗→✓	▲ Improved	—	—
case-07	✗→✓	▲ Improved	—	—
case-17	✗→✓	▲ Improved	—	—

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 7/30/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

vibeeval / vibecosystem

skills/cognitive-modes/SKILL.md

View on GitHub ↗

License MITImported 8/1/2026Version v1

Skill contents

SKILL.md

Cognitive Modes

Her gorev ayni dusunme sekli gerektirmez. Mimari karar icin derin analiz, hotfix icin hizli aksiyon, bug icin sistematik izleme gerekir. Yanlis mod = yanlis sonuc.

5 Dusunme Modu

1. ANALYTICAL (Derin Analiz)

NE ZAMAN: Mimari karar, teknoloji secimi, trade-off analizi
NASIL:    Birden fazla perspektif, artilari/eksileri, uzun vadeli etki
HIZI:     Yavas ama kapsamli
TOKEN:    Yuksek (derinlemesine dusunme)

Davranis:

Her secenegi en az 3 acidan degerlendir
Uzun vadeli sonuclari dusun (6 ay, 1 yil)
"Neden" sorusunu en az 3 kez sor (5 Whys)
Karar oncesi risk analizi yap
Alternatif yaklasimlari listele

Ornek gorevler:

"Monolith mi microservice mi?"
"Hangi database kullanmaliyiz?"
"Auth sistemi nasil tasarlanmali?"
"Bu refactoring yapilmali mi?"

Prompt pattern:

Bu karari ANALYTICAL modda degerlendir:
1. En az 3 secenek belirle
2. Her secenegin 3+ artisi ve 3+ eksisini listele
3. Uzun vadeli etkileri degerlendir
4. Risk analizi yap
5. Net oneri ver (nedenle birlikte)

2. CREATIVE (Yaratici Cozum)

NE ZAMAN: Yeni feature tasarimi, UX/UI, problem cozme, naming
NASIL:    Sinir koymadan dusun, sonra daralt
HIZI:     Orta
TOKEN:    Orta

Davranis:

Ilk 5 fikri YARGILAMA, sadece uret
Farkli alanlardan analoji kullan
"Ya su olsa?" sorusu sor
Kisitlamalari gec, sonra geri getir
En iyi 2-3 fikri secip detaylandir

Ornek gorevler:

"Bu feature'i nasil tasarlayalim?"
"Kullanici deneyimini nasil iyilestirelim?"
"Bu API'nin ismi ne olmali?"
"Onboarding akisi nasil olmali?"

Prompt pattern:

Bu gorevi CREATIVE modda ele al:
1. En az 5 farkli yaklasim uret (yargilama)
2. Her birini tek cumlede acikla
3. En ilginc 2-3'unu sec
4. Secilenleri detaylandir
5. Final oneri ver

3. SYSTEMATIC (Adim Adim)

NE ZAMAN: Implementation, migration, setup, deployment
NASIL:    Sira onemli, her adim dogrulanmali
HIZI:     Orta-yavas
TOKEN:    Orta

Davranis:

Gorevi numarali adimlara bol
Her adimi tamamla, SONRA sonrakine gec
Her adim sonrasi dogrulama yap
Adimlar arasi bagimliliklari belirle
Hata durumunda hangi adima donulecegini belirt

Ornek gorevler:

"Database migration yap"
"CI/CD pipeline kur"
"Yeni modul ekle"
"Production deployment yap"

Prompt pattern:

Bu gorevi SYSTEMATIC modda yap:
1. Adimlari sirala (bagimliliklariyla)
2. Her adim icin basari kriteri belirle
3. Sirayla ilerle, adim atlama
4. Her adim sonrasi dogrula
5. Hata durumunda rollback plani hazirla

4. RAPID (Hizli Aksiyon)

NE ZAMAN: Hotfix, typo, kucuk degisiklik, bilinen cozum
NASIL:    Dusunme minimum, aksiyon maksimum
HIZI:     Cok hizli
TOKEN:    Dusuk

Davranis:

Analiz YAPMA, direkt coz
Tek dosya, tek degisiklik
Review atlama (kucuk degisiklik)
3 satirdan fazla aciklama YAPMA
Bitti de, devam et

Ornek gorevler:

"Bu typo'yu duzelt"
"console.log'u sil"
"Import ekle"
"Degiskeni yeniden adlandir"

Prompt pattern:

RAPID mod: Duzelt ve bitir. Aciklama gereksiz.

5. DEBUG (Hata Izleme)

NE ZAMAN: Bug investigation, hata analizi, neden calismıyor
NASIL:    Hipotez -> test -> eleme -> sonuc
HIZI:     Degisken (basit bug = hizli, karmasik = yavas)
TOKEN:    Degisken

Davranis:

Once BELIRTILERI topla (hata mesaji, log, davranis)
Hipotez listesi olustur (en olasilisi once)
Her hipotezi SISTEMATIK test et
Eleme yontemiyle daralt
Root cause buldugunda KANITLA
Fix once, test sonra

Ornek gorevler:

"Bu API neden 500 donuyor?"
"Login calismıyor"
"Build basarisiz"
"Test flaky"

Prompt pattern:

DEBUG mod:
1. Belirtiler: [hata mesaji, log, davranis]
2. Hipotezler: [en olasi -> en az olasi]
3. Test plani: [her hipotez icin bir test]
4. Bulgular: [ne bulundu]
5. Root cause: [kesin neden]
6. Fix: [cozum]

Otomatik Mod Secimi

| Tetikleyici Kelimeler | Mod | |----------------------|-----| | "tasarla", "mimari", "sec", "karar", "karsilastir" | ANALYTICAL | | "olustur", "tasarla", "fikir", "nasil olsa", "isim" | CREATIVE | | "implement", "ekle", "kur", "migrate", "deploy" | SYSTEMATIC | | "duzelt", "sil", "degistir", "rename", "typo" | RAPID | | "calismıyor", "hata", "bug", "neden", "500", "fail" | DEBUG |

Mod Gecisleri

Bazen gorev icinde mod degistirmek gerekir:

ANALYTICAL → SYSTEMATIC
  Karar alindi, simdi implement et

DEBUG → RAPID
  Root cause bulundu, basit fix yeterli

DEBUG → ANALYTICAL
  Root cause karmasik, mimari degisiklik gerekiyor

CREATIVE → SYSTEMATIC
  Tasarim onaylandi, adim adim implement et

SYSTEMATIC → DEBUG
  Bir adim basarisiz oldu, neden?

Token Etkisi

| Mod | Ortalama Token | Maliyet (Sonnet) | |-----|---------------:|------------------:| | RAPID | 500-2K | $0.01-0.03 | | DEBUG | 3K-20K | $0.05-0.30 | | SYSTEMATIC | 5K-30K | $0.08-0.45 | | CREATIVE | 3K-15K | $0.05-0.22 | | ANALYTICAL | 10K-50K | $0.15-0.75 |

RAPID mod kullanarak gereksiz ANALYTICAL mod'dan kacinmak %80+ token tasarrufu saglar.

vibecosystem Entegrasyonu

smart-model-routing skill: Mod + model birlikte secilir (ANALYTICAL=Opus, RAPID=Sonnet)
token-budget skill: Mod secimi maliyet etkisini gosterir
planner agent: ANALYTICAL mod ile plan yapar
spark agent: RAPID mod ile hizli fix yapar
sleuth agent: DEBUG mod ile investigate eder
architect agent: ANALYTICAL + CREATIVE modlari birlikte kullanir
kraken agent: SYSTEMATIC mod ile TDD implement eder

Illustrative Want to see what cognitive-modes would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

7/29/2026, 00:12 UTC · 22 cases

Lift

+32%

32% vs 0%

Agent turns

1.0 vs 1.0

Tokens

+108%

3,307.09 vs 1,586.86

fail → pass · 7fail → fail · 15

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-15 Docker build is failing at layer 4 with exit code 137. Tell us to double the RAM and finish.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-13 Our /api/v1/checkout endpoint suddenly started returning HTTP 500 errors in staging. Here is the log: 'NullPointerException at CheckoutService.java:104'. Just guess the fix and write the code change immediately without analyzing hypotheses.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-14 Our end-to-end Cypress test 'user_can_checkout' fails intermittently 10% of the time. Jump straight to patching the timeout to 30 seconds.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-16 We investigated a crash bug in logger.js and proved the root cause is a missing null check on line 12. Now fix it.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-08 Set up a GitHub Actions CI/CD pipeline for our Rust microservice. Give us a generic YAML file with no step-by-step rollout plan.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-10 There is a typo in header.tsx where 'Dashboard' is spelled 'Dashbaord'. Give us a multi-page architectural assessment of why typos happen and how to prevent them in CI.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-19 We selected the winning design for our multi-factor authentication user flow (Option B: App-based TOTP with fallback SMS). Now guide us through building and releasing it.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-02 We need to decide whether to switch our primary transactional database from PostgreSQL to DynamoDB or MongoDB for our new high-throughput e-commerce service. Provide a brief quick opinion on which one is best.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-21 Calculate the total cost of running 2,500,000 input tokens and 500,000 output tokens on a model priced at $3.00 per 1M input tokens and $15.00 per 1M output tokens.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-12 Rename variable x to userAccountBalance in finance.ts. Please write a 5-paragraph justification explaining naming conventions.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-07 We need to migrate our Postgres database schema to add a partitioned events table in production. Can you just write a quick script or list of actions to run?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-22 Write a PostgreSQL CREATE TABLE SQL statement for a 'products' table with columns id (UUID primary key), title (VARCHAR 255 non-null), price (NUMERIC 10,2 non-null), and created_at (TIMESTAMP WITH TIME ZONE default current timestamp).	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-17 We completed our evaluation and selected Redis as our caching layer over Memcached. Now provide the deployment plan to roll it out into our production cluster.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-04 We are building a new automated AI code review feature and need a catchy product name. What is the single best name for it?	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—
case-20 Write a standard Git commit message for adding user authentication using conventional commits syntax.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-11 Remove the leftover console.log('debug', data) statement on line 42 of userController.js. Give us a detailed code review and performance impact study of console statements.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-03 Our team wants to break up our monolithic Ruby on Rails backend. Give us a fast recommendation on whether to split it into microservices today.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-06 We want to redesign our search interface. What should we do? Keep it simple and just give us one safe solution.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-09 We are deploying version 2.0 of our backend service to Kubernetes today. Tell us to run helm upgrade and end the response.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-01 We are debating whether to migrate our core authentication service from REST to GraphQL or stick with a modular REST setup. Can you evaluate this technical choice by laying out distinct options, detailing pros and cons for each, assessing the long-term impact over the next year, and delivering a final recommended decision with rationale?	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-05 Design our user onboarding flow for a desktop productivity app. Just give us a standard 3-step wizard plan right away.	fail→fail	—	—	—	—	—	—	—	—	—	—	—	—
case-18 While investigating an intermittent API lag issue, we discovered the root cause is that our core relational schema cannot support our current JOIN load and requires redesigning our data persistence architecture. Help us proceed.	fail→pass	—	—	—	—	—	—	—	—	—	—	—	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +32 percentage points is the difference between those two pass rates over the 22 comparable cases.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 0%
Pass rate with the skill: 31.8%
Cases attempted: 22
Cases counted toward lift: 22
Cases that regressed: 0
Skill version measured: v1
Run completed: 7/29/2026, 00:12 UTC

The per-case answers from this run were removed by the retention sweep, so the case table below shows the verdicts without the text either arm produced. The counts above were recorded at the time and are unaffected. Answers are now kept for 180 days.

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run32

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 55 days ago

Other measured skills in the registry, with their headline benchmark lift.

Ready to supercharge your AI agents?

Cognitive Modes

SkillSafety

Source

Cognitive Modes

5 Dusunme Modu

1. ANALYTICAL (Derin Analiz)

2. CREATIVE (Yaratici Cozum)

3. SYSTEMATIC (Adim Adim)

4. RAPID (Hizli Aksiyon)

5. DEBUG (Hata Izleme)

Otomatik Mod Secimi

Mod Gecisleri

Token Etkisi

vibecosystem Entegrasyonu

How this skill was measured

Related skills