79157ac077
- Observability: MLflow Tracing section complète (CLI + SDK, LLM-as-judge) - Skills §5.5: UI UX Pro Max (33.7k stars, design reasoning engine BM25) - Diagram #41: AI Fluency High vs Low paths (Anthropic research) - 3 callouts empiriques AI Fluency Index (Rev Engine 5.6×, CLAUDE.md 30%) - Hooks docs v2.1.47-2.1.50: WorktreeCreate/Remove, ConfigChange, --from-pr - CC releases tracking: v2.1.50 → v2.1.52 - 4 nouvelles évaluations ressources (AgentsView, AI Fluency, UX Pro Max) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
398 lines
21 KiB
Markdown
398 lines
21 KiB
Markdown
---
|
||
title: "Claude Code — Development Workflows Diagrams"
|
||
description: "TDD cycle, spec-first pipeline, plan-driven workflow, iterative refinement loop"
|
||
tags: [workflows, tdd, spec-first, plan-driven, iterative]
|
||
---
|
||
|
||
# Development Workflows
|
||
|
||
Proven patterns for structuring AI-assisted development sessions.
|
||
|
||
---
|
||
|
||
### TDD Red-Green-Refactor with Claude
|
||
|
||
Test-Driven Development adapted for Claude Code: write the failing test first, then ask Claude to implement only what's needed to pass it. This prevents over-engineering and ensures tests actually verify behavior.
|
||
|
||
```mermaid
|
||
flowchart TD
|
||
A([Start: New feature needed]) --> B(Write failing test<br/>with human)
|
||
B --> C(Run tests)
|
||
C --> D{Tests fail<br/>as expected?}
|
||
D -->|No: tests pass<br/>before impl!| E(Fix test — it's too weak)
|
||
E --> B
|
||
D -->|Yes: RED ✓| F(Ask Claude to implement<br/>minimal code to pass)
|
||
F --> G(Run tests again)
|
||
G --> H{Tests pass?}
|
||
H -->|No| I(Diagnose with Claude<br/>fix implementation)
|
||
I --> G
|
||
H -->|Yes: GREEN ✓| J{Code needs<br/>refactoring?}
|
||
J -->|Yes| K(Refactor with Claude)
|
||
K --> L(Run tests: still green?)
|
||
L -->|No| I
|
||
L -->|Yes: REFACTOR ✓| M{More features<br/>needed?}
|
||
J -->|No| M
|
||
M -->|Yes| B
|
||
M -->|No| N([Feature complete ✓])
|
||
|
||
style B fill:#E85D5D,color:#fff
|
||
style F fill:#E85D5D,color:#fff
|
||
style D fill:#E87E2F,color:#fff
|
||
style H fill:#E87E2F,color:#fff
|
||
style J fill:#E87E2F,color:#fff
|
||
style G fill:#7BC47F,color:#333
|
||
style K fill:#6DB3F2,color:#fff
|
||
style N fill:#7BC47F,color:#333
|
||
|
||
click A href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "TDD — New feature"
|
||
click B href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "Write failing test (RED)"
|
||
click C href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "Run tests"
|
||
click D href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "Tests fail as expected?"
|
||
click E href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "Fix test — too weak"
|
||
click F href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "Claude implements minimal code"
|
||
click G href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "Run tests again"
|
||
click H href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "Tests pass? (GREEN)"
|
||
click I href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "Diagnose with Claude"
|
||
click J href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "Code needs refactoring?"
|
||
click K href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "Refactor with Claude"
|
||
click L href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "Run tests: still green?"
|
||
click M href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "More features needed?"
|
||
click N href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/tdd-with-claude.md" "Feature complete ✓"
|
||
```
|
||
|
||
<details>
|
||
<summary>ASCII version</summary>
|
||
|
||
```
|
||
Write failing test (RED)
|
||
│
|
||
Run tests
|
||
│
|
||
Fail as expected?
|
||
├─ No → Fix test (too weak)
|
||
└─ Yes → Ask Claude: implement minimal code
|
||
│
|
||
Run tests
|
||
│
|
||
Pass? (GREEN)
|
||
├─ No → Diagnose + fix
|
||
└─ Yes → Refactor?
|
||
├─ Yes → Refactor (REFACTOR) → re-run tests
|
||
└─ No → Next feature
|
||
```
|
||
|
||
</details>
|
||
|
||
> **Source**: [TDD with Claude](../workflows/tdd-with-claude.md)
|
||
|
||
---
|
||
|
||
### Spec-First Development Pipeline
|
||
|
||
Write the specification before the code. Claude uses the spec as the single source of truth — preventing drift between what was planned and what was built.
|
||
|
||
```mermaid
|
||
flowchart LR
|
||
A([Idea / Requirement]) --> B(Write spec.md<br/>in natural language)
|
||
B --> C(Claude reviews spec<br/>for clarity + completeness)
|
||
C --> D{Spec approved<br/> by human?}
|
||
D -->|No: gaps found| E(Refine spec<br/>address gaps)
|
||
E --> C
|
||
D -->|Yes| F(Generate tests<br/>from spec)
|
||
F --> G(Generate implementation<br/>from spec + tests)
|
||
G --> H(Run test suite)
|
||
H --> I{All tests<br/>pass?}
|
||
I -->|No| J(Claude fixes<br/>implementation)
|
||
J --> H
|
||
I -->|Yes| K(Human review<br/>spec vs output)
|
||
K --> L{Matches<br/>spec?}
|
||
L -->|No| M(Update spec<br/>or implementation)
|
||
M --> K
|
||
L -->|Yes| N([Merge ✓])
|
||
|
||
style A fill:#F5E6D3,color:#333
|
||
style B fill:#6DB3F2,color:#fff
|
||
style D fill:#E87E2F,color:#fff
|
||
style I fill:#E87E2F,color:#fff
|
||
style L fill:#E87E2F,color:#fff
|
||
style N fill:#7BC47F,color:#333
|
||
|
||
click A href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Idea / Requirement"
|
||
click B href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Write spec.md"
|
||
click C href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Claude reviews spec"
|
||
click D href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Spec approved by human?"
|
||
click E href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Refine spec"
|
||
click F href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Generate tests from spec"
|
||
click G href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Generate implementation"
|
||
click H href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Run test suite"
|
||
click I href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "All tests pass?"
|
||
click J href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Claude fixes implementation"
|
||
click K href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Human review"
|
||
click L href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Matches spec?"
|
||
click M href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Update spec or implementation"
|
||
click N href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/spec-first.md" "Merge ✓"
|
||
```
|
||
|
||
<details>
|
||
<summary>ASCII version</summary>
|
||
|
||
```
|
||
Idea → Write spec.md → Claude reviews
|
||
│
|
||
Approved? ─No→ Refine spec
|
||
│ Yes
|
||
Generate tests from spec
|
||
│
|
||
Generate implementation
|
||
│
|
||
Run tests → Pass? ─No→ Claude fixes
|
||
│ Yes
|
||
Human review → Matches spec? ─No→ Fix
|
||
│ Yes
|
||
Merge ✓
|
||
```
|
||
|
||
</details>
|
||
|
||
> **Source**: [Spec-First Development](../workflows/spec-first.md)
|
||
|
||
---
|
||
|
||
### Plan-Driven Workflow with Annotation
|
||
|
||
Complex tasks benefit from plan mode: Claude explores the codebase, proposes a plan, you annotate it, then Claude executes only what was approved. Prevents surprises on large refactors.
|
||
|
||
```mermaid
|
||
flowchart TD
|
||
A([Complex task given]) --> B(Enter Plan Mode<br/>Shift+Tab × 2)
|
||
B --> C(Claude explores<br/>codebase structure)
|
||
C --> D(Claude proposes plan<br/>with file list)
|
||
D --> E(Human reviews plan)
|
||
E --> F{Plan<br/>acceptable?}
|
||
F -->|No: issues found| G(Human annotates plan<br/>marks corrections)
|
||
G --> H(Claude revises plan)
|
||
H --> E
|
||
F -->|Yes| I(Approve plan<br/>Exit Plan Mode)
|
||
I --> J(Claude executes<br/>step by step)
|
||
J --> K(Claude reports<br/>progress)
|
||
K --> L{Unexpected<br/>issue?}
|
||
L -->|Yes| M(Claude flags issue<br/>asks for guidance)
|
||
M --> F
|
||
L -->|No| N{All steps<br/>complete?}
|
||
N -->|No| J
|
||
N -->|Yes| O([Task done ✓])
|
||
|
||
style A fill:#F5E6D3,color:#333
|
||
style B fill:#6DB3F2,color:#fff
|
||
style F fill:#E87E2F,color:#fff
|
||
style L fill:#E87E2F,color:#fff
|
||
style N fill:#E87E2F,color:#fff
|
||
style G fill:#F5E6D3,color:#333
|
||
style O fill:#7BC47F,color:#333
|
||
|
||
click A href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "Complex task given"
|
||
click B href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/ultimate-guide.md#23-plan-mode" "Enter Plan Mode"
|
||
click C href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "Claude explores codebase"
|
||
click D href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "Claude proposes plan"
|
||
click E href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "Human reviews plan"
|
||
click F href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "Plan acceptable?"
|
||
click G href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "Human annotates plan"
|
||
click H href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "Claude revises plan"
|
||
click I href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/ultimate-guide.md#23-plan-mode" "Approve plan — Exit Plan Mode"
|
||
click J href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "Claude executes step by step"
|
||
click K href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "Claude reports progress"
|
||
click L href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "Unexpected issue?"
|
||
click M href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "Claude flags issue"
|
||
click N href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "All steps complete?"
|
||
click O href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/plan-driven.md" "Task done ✓"
|
||
```
|
||
|
||
<details>
|
||
<summary>ASCII version</summary>
|
||
|
||
```
|
||
Complex task
|
||
│
|
||
Plan Mode (Shift+Tab×2)
|
||
│
|
||
Claude explores codebase
|
||
│
|
||
Claude proposes plan
|
||
│
|
||
Human reviews ──No──► Annotate + Claude revises ──► re-review
|
||
│ Yes
|
||
Approve + exit plan mode
|
||
│
|
||
Claude executes step by step
|
||
│
|
||
Unexpected? ──Yes──► Flag + ask guidance
|
||
│ No
|
||
Done? ──No──► continue
|
||
│ Yes
|
||
Complete ✓
|
||
```
|
||
|
||
</details>
|
||
|
||
> **Source**: [Plan-Driven Workflow](../workflows/plan-driven.md)
|
||
|
||
---
|
||
|
||
### Iterative Refinement Loop
|
||
|
||
Output rarely hits the mark on the first try. This loop gives you a systematic way to improve results through targeted feedback rather than "make it better" vague instructions.
|
||
|
||
```mermaid
|
||
flowchart TD
|
||
A([Initial prompt]) --> B(Claude generates output)
|
||
B --> C(Evaluate output quality)
|
||
C --> D{Good<br/>enough?}
|
||
D -->|Yes| E([Done ✓])
|
||
D -->|No| F(Identify specific issue<br/>What exactly is wrong?)
|
||
F --> G{Issue type?}
|
||
G -->|Style/tone| H(Add: style constraints)
|
||
G -->|Missing info| I(Add: provide missing context)
|
||
G -->|Wrong approach| J(Add: redirect approach)
|
||
G -->|Too verbose/brief| K(Add: length constraint)
|
||
H --> L(Refine instruction)
|
||
I --> L
|
||
J --> L
|
||
K --> L
|
||
L --> M(Claude refines output)
|
||
M --> N(Compare before/after)
|
||
N --> O{Improvement<br/>detected?}
|
||
O -->|Yes| C
|
||
O -->|No| P(Different<br/>approach needed)
|
||
P --> F
|
||
|
||
style A fill:#F5E6D3,color:#333
|
||
style D fill:#E87E2F,color:#fff
|
||
style G fill:#E87E2F,color:#fff
|
||
style O fill:#E87E2F,color:#fff
|
||
style E fill:#7BC47F,color:#333
|
||
style L fill:#6DB3F2,color:#fff
|
||
style P fill:#E85D5D,color:#fff
|
||
|
||
click A href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Initial prompt"
|
||
click B href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Claude generates output"
|
||
click C href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Evaluate output quality"
|
||
click D href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Good enough?"
|
||
click E href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Done ✓"
|
||
click F href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Identify specific issue"
|
||
click G href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Issue type?"
|
||
click H href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Add style constraints"
|
||
click I href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Provide missing context"
|
||
click J href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Redirect approach"
|
||
click K href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Add length constraint"
|
||
click L href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Refine instruction"
|
||
click M href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Claude refines output"
|
||
click N href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Compare before/after"
|
||
click O href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Improvement detected?"
|
||
click P href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Different approach needed"
|
||
```
|
||
|
||
<details>
|
||
<summary>ASCII version</summary>
|
||
|
||
```
|
||
Prompt → Output → Evaluate → Good? ──Yes──► Done
|
||
│ No
|
||
Identify specific issue
|
||
│
|
||
┌──────┴──────────────┐
|
||
Style Missing Wrong Length
|
||
└──────┬──────────────┘
|
||
Refine instruction
|
||
│
|
||
Claude refines
|
||
│
|
||
Better? ──Yes──► Evaluate again
|
||
│ No
|
||
Different approach
|
||
```
|
||
|
||
</details>
|
||
|
||
> **Source**: [Iterative Refinement](../workflows/iterative-refinement.md) — Line ~347
|
||
|
||
---
|
||
|
||
### AI Fluency — High vs Low Fluency Paths
|
||
|
||
When Claude produces a polished-looking output, a cognitive bias kicks in: the more complete the output appears, the less critically most users evaluate it. This is the Artifact Paradox, documented by Anthropic across 9,830 conversations. The diagram shows what separates the 30% of high-fluency users from the 70% who accept first outputs — and the measurable difference in outcome quality.
|
||
|
||
```mermaid
|
||
flowchart TD
|
||
A([User sends request to Claude]) --> B(Claude generates output<br/>code · file · config · plan)
|
||
B --> C["⚠️ Artifact Paradox<br/>Polished output triggers<br/>cognitive acceptance bias"]
|
||
|
||
C -->|"70% of users"| D(Accept first output<br/>without critical review)
|
||
C -->|"30% of users"| E(Iterate + question<br/>define collaboration scope)
|
||
|
||
D --> D1["Fluency behaviors drop:<br/>−5.2pp gap identification<br/>−3.7pp fact-checking<br/>−3.1pp reasoning challenge"]
|
||
D1 --> D2([Silent defects · missed requirements])
|
||
|
||
E --> E1("Challenge the output:<br/>'What did you miss?<br/>What assumptions made?'")
|
||
E1 --> E2(Identify gaps<br/>Refine with full context)
|
||
E2 --> E3{Satisfied?}
|
||
E3 -->|No — iterate again| E1
|
||
E3 -->|Yes| E4([Verified, robust output ✓])
|
||
|
||
E4 --> G["Measured impact:<br/>5.6× more issue catches<br/>2.67 vs 1.33 avg behaviors<br/>Source: Anthropic AI Fluency Index, 2026"]
|
||
|
||
style A fill:#F5E6D3,color:#333
|
||
style B fill:#E87E2F,color:#fff
|
||
style C fill:#E85D5D,color:#fff
|
||
style D fill:#E85D5D,color:#fff
|
||
style D1 fill:#E85D5D,color:#fff
|
||
style D2 fill:#E85D5D,color:#fff
|
||
style E fill:#7BC47F,color:#333
|
||
style E1 fill:#6DB3F2,color:#fff
|
||
style E2 fill:#6DB3F2,color:#fff
|
||
style E3 fill:#E87E2F,color:#fff
|
||
style E4 fill:#7BC47F,color:#333
|
||
style G fill:#7BC47F,color:#333
|
||
|
||
click A href "https://www.anthropic.com/research/AI-fluency-index" "AI Fluency Index — Anthropic 2026"
|
||
click B href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/ultimate-guide.md#common-pitfalls--best-practices" "Claude generates output"
|
||
click C href "https://www.anthropic.com/research/AI-fluency-index" "Artifact Paradox — Anthropic AI Fluency Index"
|
||
click D href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/ultimate-guide.md#common-pitfalls--best-practices" "Accept without review"
|
||
click D1 href "https://www.anthropic.com/research/AI-fluency-index" "Fluency behaviors drop"
|
||
click D2 href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/ultimate-guide.md#common-pitfalls--best-practices" "Silent defects"
|
||
click E href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/ultimate-guide.md#23-plan-mode" "Iterate and question"
|
||
click E1 href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/ultimate-guide.md#rev-the-engine" "Challenge the output"
|
||
click E2 href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Identify gaps and refine"
|
||
click E3 href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Satisfied?"
|
||
click E4 href "https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/workflows/iterative-refinement.md" "Verified output"
|
||
click G href "https://www.anthropic.com/research/AI-fluency-index" "Measured impact — AI Fluency Index"
|
||
```
|
||
|
||
<details>
|
||
<summary>ASCII version</summary>
|
||
|
||
```
|
||
User request → Claude output (code · file · config · plan)
|
||
↓
|
||
⚠️ Artifact Paradox
|
||
Polished output → cognitive bias
|
||
↓
|
||
┌───────────────────┴──────────────────────┐
|
||
70% of users 30% of users
|
||
Accept without review Iterate + question
|
||
↓ ↓
|
||
Fluency behaviors drop: Challenge: "What did you miss?
|
||
−5.2pp gap identification What assumptions made?"
|
||
−3.7pp fact-checking ↓
|
||
−3.1pp reasoning challenge Identify gaps → refine
|
||
↓ ↓
|
||
Silent defects Satisfied? ──No──► iterate
|
||
↓ Yes
|
||
Verified output ✓
|
||
↓
|
||
5.6× more issue catches
|
||
2.67 vs 1.33 avg behaviors
|
||
```
|
||
|
||
</details>
|
||
|
||
> **Source**: [Anthropic AI Fluency Index](https://www.anthropic.com/research/AI-fluency-index) (Swanson et al., 2026-02-23) — [Guide section: Common Pitfalls](../ultimate-guide.md#common-pitfalls--best-practices)
|