How do I install the Evolve skill?

Run npx skillmds@latest add peopleforrester/evolve in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

What does the Evolve skill do?

Cluster related instincts into skills, commands, or agents It is listed under Coding & Dev Tools on SkillMD.

Is Evolve safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Evolve?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Evolve free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Evolve?

peopleforrester (@peopleforrester) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Evolve by peopleforrester — Agent Skill

Name: Evolve
Author: peopleforrester

File contents

/evolve

Analyze learned instincts and cluster related ones into higher-level constructs: skills, commands, or agents.

Arguments

$ARGUMENTS — Optional: domain to evolve, or "analyze" for dry-run

Process

1. Cluster Analysis

Group instincts by domain and identify clusters:

## Cluster Analysis

### Cluster: TypeScript Code Style (5 instincts, avg confidence: 0.7)
- prefer-const-assertions (0.8)
- use-discriminated-unions (0.7)
- branded-types-for-ids (0.6)
- no-enum-use-const (0.7)
- type-imports-separate (0.7)

→ Candidate for: **Skill** (coding reference)

### Cluster: Input Validation (3 instincts, avg confidence: 0.8)
- validate-inputs-with-zod (0.9)
- check-auth-on-endpoints (0.8)
- sanitize-html-output (0.7)

→ Candidate for: **Rule** (always-enforce constraint)

### Cluster: Database Review (4 instincts, avg confidence: 0.6)
- check-n-plus-one (0.7)
- use-indexes-for-where (0.6)
- paginate-list-endpoints (0.5)
- use-transactions (0.6)

→ Candidate for: **Agent prompt enhancement** (add to database-reviewer)

2. Evolution Thresholds

Target	Min Instincts	Min Avg Confidence	Criteria
Skill	3+	0.5	Related workflow patterns
Rule	3+	0.7	Hard constraints (always/never)
Command	3+	0.5	User-invoked workflow
Agent enhancement	4+	0.5	Domain-specific expertise

3. Generation

For approved clusters, generate the appropriate construct:

Skill generation:

---
name: typescript-style-evolved
description: |
  Evolved from 5 learned instincts about TypeScript coding style.
  Covers const assertions, discriminated unions, branded types.
---
# TypeScript Style Patterns (Learned)
[Content synthesized from instinct evidence]

4. Output Location

Generated constructs are saved to:

~/.claude/evolved/
  skills/typescript-style-evolved/SKILL.md
  rules/input-validation-evolved.md

5. Report

## Evolution Report

| Cluster | Size | Confidence | Evolved To | Status |
|---------|------|-----------|-----------|--------|
| TS Code Style | 5 | 0.7 | Skill | Created |
| Input Validation | 3 | 0.8 | Rule | Created |
| Database Review | 4 | 0.6 | Agent enhancement | Suggested |

### Next Steps
- Review generated files in `~/.claude/evolved/`
- Copy approved files to appropriate directories
- Run `/instinct-status` to see remaining unclustered instincts

Dry Run

Use /evolve analyze to see cluster analysis without generating files.

---name: evolvedescription: Cluster related instincts into skills, commands, or agents---# /evolveAnalyze learned instincts and cluster related ones into higher-levelconstructs: skills, commands, or agents.## Arguments- `$ARGUMENTS` — Optional: domain to evolve, or "analyze" for dry-run## Process### 1. Cluster AnalysisGroup instincts by domain and identify clusters:```markdown## Cluster Analysis### Cluster: TypeScript Code Style (5 instincts, avg confidence: 0.7)- prefer-const-assertions (0.8)- use-discriminated-unions (0.7)- branded-types-for-ids (0.6)- no-enum-use-const (0.7)- type-imports-separate (0.7)→ Candidate for: **Skill** (coding reference)### Cluster: Input Validation (3 instincts, avg confidence: 0.8)- validate-inputs-with-zod (0.9)- check-auth-on-endpoints (0.8)- sanitize-html-output (0.7)→ Candidate for: **Rule** (always-enforce constraint)### Cluster: Database Review (4 instincts, avg confidence: 0.6)- check-n-plus-one (0.7)- use-indexes-for-where (0.6)- paginate-list-endpoints (0.5)- use-transactions (0.6)→ Candidate for: **Agent prompt enhancement** (add to database-reviewer)```### 2. Evolution Thresholds| Target | Min Instincts | Min Avg Confidence | Criteria ||--------|--------------|-------------------|----------|| Skill | 3+ | 0.5 | Related workflow patterns || Rule | 3+ | 0.7 | Hard constraints (always/never) || Command | 3+ | 0.5 | User-invoked workflow || Agent enhancement | 4+ | 0.5 | Domain-specific expertise |### 3. GenerationFor approved clusters, generate the appropriate construct:**Skill generation:**```markdown---name: typescript-style-evolveddescription: | Evolved from 5 learned instincts about TypeScript coding style. Covers const assertions, discriminated unions, branded types.---# TypeScript Style Patterns (Learned)[Content synthesized from instinct evidence]```### 4. Output LocationGenerated constructs are saved to:```~/.claude/evolved/ skills/typescript-style-evolved/SKILL.md rules/input-validation-evolved.md```### 5. Report```markdown## Evolution Report| Cluster | Size | Confidence | Evolved To | Status ||---------|------|-----------|-----------|--------|| TS Code Style | 5 | 0.7 | Skill | Created || Input Validation | 3 | 0.8 | Rule | Created || Database Review | 4 | 0.6 | Agent enhancement | Suggested |### Next Steps- Review generated files in `~/.claude/evolved/`- Copy approved files to appropriate directories- Run `/instinct-status` to see remaining unclustered instincts```## Dry RunUse `/evolve analyze` to see cluster analysis without generating files.

Frequently asked questions

Evolve