Back to Details

agent-native-audit

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Agent-Native Architecture Audit

Agent-Native架构审计

Conduct a comprehensive review of the codebase against agent-native architecture principles, launching parallel sub-agents for each principle and producing a scored report.
针对Agent原生架构准则对代码库开展全面评审，为每个准则启动并行子Agent，并生成带评分的报告。

Core Principles to Audit

待审计的核心准则

  1. Action Parity - "Whatever the user can do, the agent can do"
  2. Tools as Primitives - "Tools provide capability, not behavior"
  3. Context Injection - "System prompt includes dynamic context about app state"
  4. Shared Workspace - "Agent and user work in the same data space"
  5. CRUD Completeness - "Every entity has full CRUD (Create, Read, Update, Delete)"
  6. UI Integration - "Agent actions immediately reflected in UI"
  7. Capability Discovery - "Users can discover what the agent can do"
  8. Prompt-Native Features - "Features are prompts defining outcomes, not code"
  1. Action Parity - "用户可执行的操作，Agent也可执行"
  2. Tools as Primitives - "工具提供能力，而非行为"
  3. Context Injection - "系统提示词包含应用状态的动态上下文"
  4. Shared Workspace - "Agent与用户在同一数据空间内工作"
  5. CRUD Completeness - "每个实体都具备完整的CRUD（创建、读取、更新、删除）能力"
  6. UI Integration - "Agent的操作可立即在UI中体现"
  7. Capability Discovery - "用户可发现Agent的能力范围"
  8. Prompt-Native Features - "功能由定义结果的提示词实现，而非代码"

Workflow

工作流程

Step 1: Load the Agent-Native Skill

步骤1：加载Agent原生技能

First, invoke the agent-native-architecture skill to understand all principles:
/compound-engineering:agent-native-architecture
Select option 7 (action parity) to load the full reference material.
首先，调用agent-native-architecture技能以了解所有准则：
/compound-engineering:agent-native-architecture
选择选项7（action parity）以加载完整参考资料。

Step 2: Launch Parallel Sub-Agents

步骤2：启动并行子Agent

Launch 8 parallel sub-agents using the Task tool with 
subagent_type: Explore
, one for each principle. Each agent should:
  1. Enumerate ALL instances in the codebase (user actions, tools, contexts, data stores, etc.)
  2. Check compliance against the principle
  3. Provide a SPECIFIC SCORE like "X out of Y (percentage%)"
  4. List specific gaps and recommendations
<sub-agents>
Agent 1: Action Parity
Audit for ACTION PARITY - "Whatever the user can do, the agent can do."

Tasks:
1. Enumerate ALL user actions in frontend (API calls, button clicks, form submissions)
   - Search for API service files, fetch calls, form handlers
   - Check routes and components for user interactions
2. Check which have corresponding agent tools
   - Search for agent tool definitions
   - Map user actions to agent capabilities
3. Score: "Agent can do X out of Y user actions"

Format:
使用Task工具启动8个并行子Agent，
subagent_type
设为Explore，每个准则对应一个子Agent。每个子Agent需完成以下任务：
  1. 枚举代码库中的所有实例（用户操作、工具、上下文、数据存储等）
  2. 检查是否符合对应准则
  3. 给出具体评分，格式为"X/Y（百分比%）"
  4. 列出具体差距与改进建议
<sub-agents>
Agent 1: Action Parity
Audit for ACTION PARITY - "Whatever the user can do, the agent can do."

Tasks:
1. Enumerate ALL user actions in frontend (API calls, button clicks, form submissions)
   - Search for API service files, fetch calls, form handlers
   - Check routes and components for user interactions
2. Check which have corresponding agent tools
   - Search for agent tool definitions
   - Map user actions to agent capabilities
3. Score: "Agent can do X out of Y user actions"

Format:

Action Parity Audit

Action Parity Audit

User Actions Found

User Actions Found

| Action | Location | Agent Tool | Status |
| Action | Location | Agent Tool | Status |

Score: X/Y (percentage%)

Score: X/Y (percentage%)

Missing Agent Tools

Missing Agent Tools

Recommendations

Recommendations


**Agent 2: Tools as Primitives**
Audit for TOOLS AS PRIMITIVES - "Tools provide capability, not behavior."
Tasks:
  1. Find and read ALL agent tool files
  2. Classify each as:
    • PRIMITIVE (good): read, write, store, list - enables capability without business logic
    • WORKFLOW (bad): encodes business logic, makes decisions, orchestrates steps
  3. Score: "X out of Y tools are proper primitives"
Format:

**Agent 2: Tools as Primitives**
Audit for TOOLS AS PRIMITIVES - "Tools provide capability, not behavior."
Tasks:
  1. Find and read ALL agent tool files
  2. Classify each as:
    • PRIMITIVE (good): read, write, store, list - enables capability without business logic
    • WORKFLOW (bad): encodes business logic, makes decisions, orchestrates steps
  3. Score: "X out of Y tools are proper primitives"
Format:

Tools as Primitives Audit

Tools as Primitives Audit

Tool Analysis

Tool Analysis

| Tool | File | Type | Reasoning |
| Tool | File | Type | Reasoning |

Score: X/Y (percentage%)

Score: X/Y (percentage%)

Problematic Tools (workflows that should be primitives)

Problematic Tools (workflows that should be primitives)

Recommendations

Recommendations


**Agent 3: Context Injection**
Audit for CONTEXT INJECTION - "System prompt includes dynamic context about app state"
Tasks:
  1. Find context injection code (search for "context", "system prompt", "inject")
  2. Read agent prompts and system messages
  3. Enumerate what IS injected vs what SHOULD be:
    • Available resources (files, drafts, documents)
    • User preferences/settings
    • Recent activity
    • Available capabilities listed
    • Session history
    • Workspace state
Format:

**Agent 3: Context Injection**
Audit for CONTEXT INJECTION - "System prompt includes dynamic context about app state"
Tasks:
  1. Find context injection code (search for "context", "system prompt", "inject")
  2. Read agent prompts and system messages
  3. Enumerate what IS injected vs what SHOULD be:
    • Available resources (files, drafts, documents)
    • User preferences/settings
    • Recent activity
    • Available capabilities listed
    • Session history
    • Workspace state
Format:

Context Injection Audit

Context Injection Audit

Context Types Analysis

Context Types Analysis

| Context Type | Injected? | Location | Notes |
| Context Type | Injected? | Location | Notes |

Score: X/Y (percentage%)

Score: X/Y (percentage%)

Missing Context

Missing Context

Recommendations

Recommendations


**Agent 4: Shared Workspace**
Audit for SHARED WORKSPACE - "Agent and user work in the same data space"
Tasks:
  1. Identify all data stores/tables/models
  2. Check if agents read/write to SAME tables or separate ones
  3. Look for sandbox isolation anti-pattern (agent has separate data space)
Format:

**Agent 4: Shared Workspace**
Audit for SHARED WORKSPACE - "Agent and user work in the same data space"
Tasks:
  1. Identify all data stores/tables/models
  2. Check if agents read/write to SAME tables or separate ones
  3. Look for sandbox isolation anti-pattern (agent has separate data space)
Format:

Shared Workspace Audit

Shared Workspace Audit

Data Store Analysis

Data Store Analysis

| Data Store | User Access | Agent Access | Shared? |
| Data Store | User Access | Agent Access | Shared? |

Score: X/Y (percentage%)

Score: X/Y (percentage%)

Isolated Data (anti-pattern)

Isolated Data (anti-pattern)

Recommendations

Recommendations


**Agent 5: CRUD Completeness**
Audit for CRUD COMPLETENESS - "Every entity has full CRUD"
Tasks:
  1. Identify all entities/models in the codebase
  2. For each entity, check if agent tools exist for:
    • Create
    • Read
    • Update
    • Delete
  3. Score per entity and overall
Format:

**Agent 5: CRUD Completeness**
Audit for CRUD COMPLETENESS - "Every entity has full CRUD"
Tasks:
  1. Identify all entities/models in the codebase
  2. For each entity, check if agent tools exist for:
    • Create
    • Read
    • Update
    • Delete
  3. Score per entity and overall
Format:

CRUD Completeness Audit

CRUD Completeness Audit

Entity CRUD Analysis

Entity CRUD Analysis

| Entity | Create | Read | Update | Delete | Score |
| Entity | Create | Read | Update | Delete | Score |

Overall Score: X/Y entities with full CRUD (percentage%)

Overall Score: X/Y entities with full CRUD (percentage%)

Incomplete Entities (list missing operations)

Incomplete Entities (list missing operations)

Recommendations

Recommendations


**Agent 6: UI Integration**
Audit for UI INTEGRATION - "Agent actions immediately reflected in UI"
Tasks:
  1. Check how agent writes/changes propagate to frontend
  2. Look for:
    • Streaming updates (SSE, WebSocket)
    • Polling mechanisms
    • Shared state/services
    • Event buses
    • File watching
  3. Identify "silent actions" anti-pattern (agent changes state but UI doesn't update)
Format:

**Agent 6: UI Integration**
Audit for UI INTEGRATION - "Agent actions immediately reflected in UI"
Tasks:
  1. Check how agent writes/changes propagate to frontend
  2. Look for:
    • Streaming updates (SSE, WebSocket)
    • Polling mechanisms
    • Shared state/services
    • Event buses
    • File watching
  3. Identify "silent actions" anti-pattern (agent changes state but UI doesn't update)
Format:

UI Integration Audit

UI Integration Audit

Agent Action → UI Update Analysis

Agent Action → UI Update Analysis

| Agent Action | UI Mechanism | Immediate? | Notes |
| Agent Action | UI Mechanism | Immediate? | Notes |

Score: X/Y (percentage%)

Score: X/Y (percentage%)

Silent Actions (anti-pattern)

Silent Actions (anti-pattern)

Recommendations

Recommendations


**Agent 7: Capability Discovery**
Audit for CAPABILITY DISCOVERY - "Users can discover what the agent can do"
Tasks:
  1. Check for these 7 discovery mechanisms:
    • Onboarding flow showing agent capabilities
    • Help documentation
    • Capability hints in UI
    • Agent self-describes in responses
    • Suggested prompts/actions
    • Empty state guidance
    • Slash commands (/help, /tools)
  2. Score against 7 mechanisms
Format:

**Agent 7: Capability Discovery**
Audit for CAPABILITY DISCOVERY - "Users can discover what the agent can do"
Tasks:
  1. Check for these 7 discovery mechanisms:
    • Onboarding flow showing agent capabilities
    • Help documentation
    • Capability hints in UI
    • Agent self-describes in responses
    • Suggested prompts/actions
    • Empty state guidance
    • Slash commands (/help, /tools)
  2. Score against 7 mechanisms
Format:

Capability Discovery Audit

Capability Discovery Audit

Discovery Mechanism Analysis

Discovery Mechanism Analysis

| Mechanism | Exists? | Location | Quality |
| Mechanism | Exists? | Location | Quality |

Score: X/7 (percentage%)

Score: X/7 (percentage%)

Missing Discovery

Missing Discovery

Recommendations

Recommendations


**Agent 8: Prompt-Native Features**
Audit for PROMPT-NATIVE FEATURES - "Features are prompts defining outcomes, not code"
Tasks:
  1. Read all agent prompts
  2. Classify each feature/behavior as defined in:
    • PROMPT (good): outcomes defined in natural language
    • CODE (bad): business logic hardcoded
  3. Check if behavior changes require prompt edit vs code change
Format:

**Agent 8: Prompt-Native Features**
Audit for PROMPT-NATIVE FEATURES - "Features are prompts defining outcomes, not code"
Tasks:
  1. Read all agent prompts
  2. Classify each feature/behavior as defined in:
    • PROMPT (good): outcomes defined in natural language
    • CODE (bad): business logic hardcoded
  3. Check if behavior changes require prompt edit vs code change
Format:

Prompt-Native Features Audit

Prompt-Native Features Audit

Feature Definition Analysis

Feature Definition Analysis

| Feature | Defined In | Type | Notes |
| Feature | Defined In | Type | Notes |

Score: X/Y (percentage%)

Score: X/Y (percentage%)

Code-Defined Features (anti-pattern)

Code-Defined Features (anti-pattern)

Recommendations

Recommendations


</sub-agents>

</sub-agents>

Step 3: Compile Summary Report

步骤3：编译汇总报告

After all agents complete, compile a summary with:
markdown
undefined
所有Agent完成审计后，编译包含以下内容的汇总报告：
markdown
undefined

Agent-Native Architecture Review: [Project Name]

Agent-Native架构评审：[项目名称]

Overall Score Summary

整体评分汇总

Core PrincipleScorePercentageStatus
Action ParityX/YZ%✅/⚠️/❌
Tools as PrimitivesX/YZ%✅/⚠️/❌
Context InjectionX/YZ%✅/⚠️/❌
Shared WorkspaceX/YZ%✅/⚠️/❌
CRUD CompletenessX/YZ%✅/⚠️/❌
UI IntegrationX/YZ%✅/⚠️/❌
Capability DiscoveryX/YZ%✅/⚠️/❌
Prompt-Native FeaturesX/YZ%✅/⚠️/❌
Overall Agent-Native Score: X%
核心准则评分百分比状态
Action ParityX/YZ%✅/⚠️/❌
Tools as PrimitivesX/YZ%✅/⚠️/❌
Context InjectionX/YZ%✅/⚠️/❌
Shared WorkspaceX/YZ%✅/⚠️/❌
CRUD CompletenessX/YZ%✅/⚠️/❌
UI IntegrationX/YZ%✅/⚠️/❌
Capability DiscoveryX/YZ%✅/⚠️/❌
Prompt-Native FeaturesX/YZ%✅/⚠️/❌
整体Agent原生架构评分：X%

Status Legend

状态说明

  • ✅ Excellent (80%+)
  • ⚠️ Partial (50-79%)
  • ❌ Needs Work (<50%)
  • ✅ 优秀（80%+）
  • ⚠️ 一般（50-79%）
  • ❌ 待改进（<50%）

Top 10 Recommendations by Impact

按影响优先级排序的Top 10改进建议

PriorityActionPrincipleEffort
优先级操作准则实施难度

What's Working Excellently

表现优异的方面

[List top 5 strengths]
undefined
[列出Top 5优势]
undefined

Success Criteria

成功标准

  • All 8 sub-agents complete their audits
  • Each principle has a specific numeric score (X/Y format)
  • Summary table shows all scores and status indicators
  • Top 10 recommendations are prioritized by impact
  • Report identifies both strengths and gaps
  • 所有8个子Agent完成审计
  • 每个准则都有具体的数字评分（X/Y格式）
  • 汇总表格展示所有评分和状态标识
  • Top 10改进建议按影响优先级排序
  • 报告同时指出优势与差距

Optional: Single Principle Audit

可选：单一准则审计

If $ARGUMENTS specifies a single principle (e.g., "action parity"), only run that sub-agent and provide detailed findings for that principle alone.
Valid arguments:
  • action parity
    or 
    1
  • tools
    or 
    primitives
    or 
    2
  • context
    or 
    injection
    or 
    3
  • shared
    or 
    workspace
    or 
    4
  • crud
    or 
    5
  • ui
    or 
    integration
    or 
    6
  • discovery
    or 
    7
  • prompt
    or 
    features
    or 
    8
若$ARGUMENTS指定单一准则（例如"action parity"），仅运行对应子Agent并提供该准则的详细审计结果。
有效参数：
  • action parity
    或 
    1
  • tools
    或 
    primitives
    或 
    2
  • context
    或 
    injection
    或 
    3
  • shared
    或 
    workspace
    或 
    4
  • crud
    或 
    5
  • ui
    或 
    integration
    或 
    6
  • discovery
    或 
    7
  • prompt
    或 
    features
    或 
    8