Assess agentic layer maturity using the 12-grade classification system (Class 1-3). Use when evaluating codebase readiness, identifying next upgrade steps, or tracking progress toward the Codebase Singularity.
中文简介正在准备中。当前显示原文说明。
查看安装方法安装前可以查看交给 AI 代理的指令内容。
Assess agentic layer maturity using the complete 12-grade classification system from TAC Lesson 14.
.claude/ directoryThree classes with 12 total grades:
| Grade | Component | Indicator |
|---|---|---|
| 1 | Memory Files | CLAUDE.md exists with guidance |
| 2 | Sub-Agents | Task agents used for parallelization |
| 3 | Skills/MCPs | Custom skills or MCP integrations |
| 4 | Closed-Loops | Self-validating prompts |
| 5 | Templates | Bug/feature/chore classification |
| 6 | Prompt Chains | Multi-step composite workflows |
| 7 | Agent Experts | Expertise files with self-improve |
| Grade | Component | Indicator |
|---|---|---|
| 1 | Webhooks | External triggers (PITER framework) |
| 2 | ADWs | AI Developer Workflows running |
| Grade | Component | Indicator |
|---|---|---|
| 1 | Orchestrator | Meta-agent managing fleet |
| 2 | Orchestrator Workflows | Human-orchestrator interaction |
| 3 | ADWs + Orchestrator | Full autonomous execution |
Check for indicators of each grade:
# Grade 1: Memory files
ls .claude/ CLAUDE.md
# Grade 2: Sub-agents
ls .claude/agents/
# Grade 3: Skills
ls .claude/skills/ || ls -d */skills/ 2>/dev/null
# Grade 4: Closed-loop patterns
grep -r "validation" .claude/commands/
grep -r "retry" .claude/commands/
# Grade 5: Templates
ls .claude/commands/ | grep -E "(chore|bug|feature)"
# Grade 6: Prompt chains
grep -r "Step 1" .claude/commands/
grep -r "Then execute" .claude/commands/
# Grade 7: Agent experts
ls .claude/commands/experts/ 2>/dev/null
find . -name "expertise.yaml"
# Grade 8 (Class 2 G1): Webhooks
find . -name "*webhook*" -o -name "*trigger*"
# Grade 9 (Class 2 G2): ADWs
ls adws/ 2>/dev/null
# Grade 10-12 (Class 3): Orchestrator
find . -name "*orchestrator*"
For each grade, determine status:
| Status | Meaning |
|---|---|
| ✅ Complete | Fully implemented and used |
| 🔶 Partial | Some elements present |
| ❌ Missing | Not implemented |
Your level = highest consecutive completed grade
Example:
Result: Class 1 Grade 4 (solid), targeting Grade 5
Recommend specific actions for next grade:
| Current | Next Step |
|---|---|
| Grade 1 | Add Task agents for parallelization |
| Grade 2 | Create custom skills or MCP |
| Grade 3 | Add validation loops to prompts |
| Grade 4 | Implement issue classification templates |
| Grade 5 | Chain prompts into workflows |
| Grade 6 | Build first agent expert |
| Grade 7 | Set up external triggers |
| C2G1 | Implement AI Developer Workflows |
| C2G2 | Build orchestrator agent |
| C3G1 | Add human-orchestrator workflows |
| C3G2 | Connect orchestrator to ADWs |
## Agentic Layer Assessment Report
**Codebase:** [project name]
**Date:** [assessment date]
**Assessed by:** [model]
### Classification Summary
**Current Level:** Class [1/2/3] Grade [1-7/1-2/1-3]
**Maturity Score:** [X]/12 grades achieved
### Grade-by-Grade Assessment
| Grade | Component | Status | Evidence |
| --- | --- | --- | --- |
| C1G1 | Memory Files | ✅/🔶/❌ | [what was found] |
| C1G2 | Sub-Agents | ✅/🔶/❌ | [what was found] |
...
### Strengths
- [What's working well]
### Gaps
- [What's missing or weak]
### Recommended Next Steps
1. **Priority 1:** [Most impactful improvement]
2. **Priority 2:** [Second priority]
3. **Priority 3:** [Third priority]
### Path to Class 3
[Roadmap of remaining grades to achieve]
.claude/ directory structure"Your agentic layer should be specialized to fit and wrap your codebase. Don't focus on reuse, focus on making these prompts great for that one codebase."
Each grade builds on the previous. Skip a grade and the foundation becomes unstable.
| Anti-Pattern | Problem | Solution |
|---|---|---|
| Skipping grades | Missing foundation | Build progressively |
| Over-engineering early | Complexity before value | Start with Grade 1-2 |
| Generic layers | Don't fit codebase | Specialize for your project |
| Assessment without action | No improvement | Prioritize next step |
Date: 2026-01-01 Model: claude-opus-4-5-20251101
还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问，或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
[00] TnR ABSOLUTE RIGOR — Systemic Verification Doctrine. Forced cognitive exertion for high-stakes investigations. Mandates absolute deconstruction, mathematical proof chains, and ≥95% derived certainty. Zero-tolerance for heuristics or sentiment-based reasoning. Shortcuts are defined as intentional hallucinations.
中文简介正在准备中。当前显示原文说明。
Configures 2-Agent workflow between PM and implementation roles. Use when user mentions 2-Agent, 2エージェント, PM連携設定, Cursor設定, Cursor連携, 2-agent運用. Do NOT load for: 単独運用, ワークフロー実行, ハンドオフ処理.
中文简介正在准备中。当前显示原文说明。
Appends relevant hashtags to journal entries based on content analysis. Intended for LLM execution.
中文简介正在准备中。当前显示原文说明。
Add a new AI coding agent to Agent Sandbox. Creates all required files (Dockerfile, templates, CI, docs) and wires the agent into the CLI, proxy, and build system.
中文简介正在准备中。当前显示原文说明。
Add a new property to the AI agents database. Use when the user wants to add, create, or introduce a new column, property, field, or feature to track across all agents in the comparison matrix. Handles all four required steps - database updates, groups.json, table display, and GitHub issue templates.
中文简介正在准备中。当前显示原文说明。
Add new Agent templates to the meta-agent-skills framework.
中文简介正在准备中。当前显示原文说明。