agentsclimarketplace

Evolution log

Skill aman-bhandari/claude-code-agent-skills-framework/.claude/skills/evolution-log

Research scaffold for AI engineering with Claude Code. 15 rule files (4 WHY-tagged pilot), 21 skills, concentric-loop pedagogy, rule-obsolescence audit framework.

Install
npx -y skills add aman-bhandari/claude-code-agent-skills-framework --skill evolution-log

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Activate at the end of every Topic completion. Generates a System Evolution Log comparing code complexity, identifying meta-patterns mastered, documenting intentional agent stress-tests, and evaluating the student's growth trajectory. Invoke with /evolution.

SKILL.md

3.4 KB, as published. Nobody here has run it

System Evolution Log (Telemetry of Mastery)

When This Fires

  • At the end of every Topic (mandatory)
  • At milestone completion
  • Manually via /evolution

The Evolution Log Format

Written to evolution/topic-N-evolution.md:

# System Evolution Log: Topic N → Topic N+1

**Date:** YYYY-MM-DD
**Sessions covered:** [session range]
**Time invested:** [hours]

## Code Complexity Comparison

### Topic N-1 Code
- Lines of code: [count]
- Number of classes/functions: [count]
- Test count: [count]
- Design patterns used: [list]
- Abstraction depth: [max nesting / composition depth]

### Topic N Code
- Lines of code: [count]
- Number of classes/functions: [count]
- Test count: [count]
- Design patterns used: [list]
- Abstraction depth: [max nesting / composition depth]

### Delta
- Complexity growth: [quantified]
- New patterns introduced: [list with first-use date]
- Patterns that became automatic: [list -- things the student does without thinking]

## Meta-Patterns Mastered

Identify 3-5 patterns the student has internalized across topics:

1. **[Pattern Name]** (e.g., "Dependency Inversion in AI Pipelines")
   - First encountered: [topic/session]
   - Became automatic: [topic/session]
   - Evidence: [specific code example where student applied it without prompting]

2. **[Pattern Name]**
   ...

## Agent Stress-Test Log

### Intentional Mistakes Used
Document any deliberate mistakes introduced to test the agent's coaching:
- What was the mistake?
- Did the agent catch it?
- How did the agent's recovery logic work?
- What does this reveal about the agent's reliability?

### Agent Coaching Quality
- Concepts where the agent excelled: [list]
- Concepts where the agent struggled: [list]
- Teaching approach changes made during this topic: [list]
- Model version used: [Claude model ID]

## Trajectory Evaluation

### Assessment Scores: Start vs End of Topic
| Dimension | Start | End | Delta | Evidence |
|-----------|-------|-----|-------|----------|
| Python Engineering | | | | |
| Problem Decomposition | | | | |
| Mathematical Thinking | | | | |
| ML/DL Understanding | | | | |
| System Design | | | | |
| Human-AI Collaboration | | | | |
| Engineering Craft | | | | |
| Evaluation Thinking | | | | |
| Communication | | | | |

### Velocity
- Exercises per session (average): [number]
- Time per exercise (average): [minutes]
- Trend: [accelerating / steady / decelerating]

### The "Click Moments"
List the 2-3 moments during this topic where understanding visibly shifted:
1. [What clicked, when, what triggered it]

### Open Gaps Carried Forward
- [Concept] -- deferred to Topic N+1 because [reason]

Why This Matters

At 18 months, the evolution/ directory contains a longitudinal record of growth that no interview can replicate. A recruiter reads topic-4-evolution.md and sees:

  • Quantified complexity growth
  • Named meta-patterns (not just "I learned stuff")
  • Agent reliability data (the student EVALUATES the system that teaches them)
  • Honest trajectory with click moments and open gaps

This is the "Telemetry of Mastery" -- the black box recorder of an engineer's development.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.