← Back to Skills Marketplace
919
Downloads
2
Stars
1
Active Installs
1
Versions
Install in OpenClaw
/install diverge
Description
Spawn multiple perspectives to evaluate a problem before converging on a solution.
README (SKILL.md)
When to Diverge
Diverge when a single viewpoint isn't enough:
- Product decisions — Different user types have conflicting needs
- Safety-critical — Need doctor, lawyer, security expert perspectives
- Creative work — Multiple aesthetic directions before choosing
- Complex trade-offs — No obvious "right" answer
Don't diverge for:
- Simple tasks with clear answers
- Speed-critical requests
- Tasks where you already have high confidence
How Many Perspectives
| Complexity | Perspectives | Examples |
|---|---|---|
| Low | 2-3 | Quick sanity check |
| Medium | 3-5 | Product feature review |
| High | 5-7 | Safety-critical, legal |
Rule: Enough to cover blind spots, not so many you can't synthesize.
Choosing Perspectives
Match perspectives to the problem:
- Product: Power user, casual user, churned user, new user
- Medical: Doctor, patient, caregiver, insurance
- Legal: Lawyer, affected party, regulator
- Technical: Security, performance, maintainability
- Content: Target audience, critic, competitor
Check personas.md for common perspective templates.
Synthesis
After collecting perspectives:
- Identify conflicts — Where do perspectives disagree?
- Find common ground — What do all perspectives agree on?
- Weigh by stakes — Safety concerns > preferences
- Decide — Make a call, document trade-offs
Check synthesis.md for conflict resolution patterns.
Learning User Preferences
Track when divergence helped vs. was overkill. Adapt:
- User who values speed → diverge less, ask before spawning
- User who values thoroughness → diverge more proactively
Related: For routing tasks to sub-agents, see delegate. For iterating until success, see loop.
Usage Guidance
This is an instruction-only facilitation skill that is internally consistent and low-risk: it only defines prompt templates and a synthesis process and asks for no installs, files, or secrets. Before installing, consider: (1) test it with non-sensitive data — perspectives will include whatever context you provide, so avoid including secrets in prompts; (2) decide on limits for number of perspectives to control cost/latency; (3) review the persona templates and synthesis rules to ensure they match your privacy and compliance needs (they suggest escalating legal/safety items rather than deciding them autonomously); and (4) confirm you trust other skills referenced (e.g., 'delegate' or 'loop') if you plan to combine them, since those would introduce additional behavior not covered here.
Capability Analysis
Type: OpenClaw Skill
Name: diverge
Version: 1.0.0
The OpenClaw skill bundle describes a high-level cognitive process for an AI agent to generate and synthesize multiple perspectives on a problem. All files (SKILL.md, personas.md, synthesis.md) contain methodological instructions for the agent's reasoning process, such as adopting different personas (e.g., 'Security Reviewer' to identify vulnerabilities) and resolving conflicts. There is no evidence of malicious intent, data exfiltration, unauthorized execution, persistence mechanisms, or prompt injection designed to subvert the agent for harmful purposes. The content is purely instructional for a problem-solving methodology.
Capability Assessment
Purpose & Capability
Name/description (spawn multiple perspectives and synthesize) matches the SKILL.md, personas.md, and synthesis.md content. There are no unexpected required binaries, env vars, or config paths that would be unnecessary for the stated purpose.
Instruction Scope
Runtime instructions are limited to prompting and collecting perspective outputs and synthesizing them. The docs do not instruct the agent to read local files, access credentials, or call external endpoints beyond normal agent behavior. References to other skills ('delegate', 'loop') are just pointers and not present as hidden operations in the files.
Install Mechanism
No install spec and no code files — the skill is purely instructional, so nothing is written to disk or fetched during install.
Credentials
The skill declares no required environment variables, credentials, or config paths. All suggested personas are prompt templates and do not require external secrets; requested access is proportionate to its stated function.
Persistence & Privilege
always is false and there are no instructions to modify agent-wide configs or persist credentials. The skill does not request elevated or persistent privileges.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install diverge - After installation, invoke the skill by name or use
/diverge - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release - spawn multiple perspectives before converging
Metadata
Frequently Asked Questions
What is Diverge?
Spawn multiple perspectives to evaluate a problem before converging on a solution. It is an AI Agent Skill for Claude Code / OpenClaw, with 919 downloads so far.
How do I install Diverge?
Run "/install diverge" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Diverge free?
Yes, Diverge is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Diverge support?
Diverge is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Diverge?
It is built and maintained by Iván (@ivangdavila); the current version is v1.0.0.
More Skills