Safety & Boundaries

• Always ask user before creating new skills or scripts
• Never auto-execute created scripts without confirmation
• Research only — do NOT implement features that require core OpenClaw changes
• Report findings to user, let them decide what to keep

Study and Port — Research &移植框架优点

"Don't just use other AI frameworks — learn from them."

When you discover an AI framework or tool with features worth learning, execute this skill.

Trigger Conditions

• User mentions a new AI framework or tool
• Discover features in other frameworks that OpenClaw doesn't have
• Want to learn from other AI assistants' strengths

Workflow

Step 1 — Quick Overview (10 min)

Search endpoint: Use MiniMax web search (configured in TOOLS.md)

Use search to understand:

Template A (General Framework):

"[framework name] features capabilities 2026"
"[framework name] vs OpenAI agent differences"
"[framework name] Hermes Agent comparison"

Template B (AI Coding Assistants):

"[framework name] features capabilities 2026"
"[framework name] self-improving memory workflow"

Template C (Chinese Frameworks):

"[framework name] 特点 功能 优势"
"[framework name] 和 OpenClaw 对比"

Collect:

• Framework name and version
• Core features (3-5)
• Design philosophy
• How it differs from OpenClaw

Step 2 — Deep Dive (as needed)

For each core feature:

"[framework name]" "[specific feature]" "how it works"
"[framework name]" "implementation" "architecture"

Extract the 3-5 most valuable learnings.

Step 3 — Feasibility Evaluation

For each feature, answer:

Portability Rating:

• 🌟 Portable (skill) — implementable as SKILL.md ± scripts
• 🔧 Partially Portable — core part doable, limitations exist
• 🏗️ Requires Core Change — needs OpenClaw source modification
• ⏳ Wait & Watch — tech not mature enough

Step 4 — Create Skill (if portable)

User confirmation required: Before creating any new skill or script, ALWAYS ask user:

• "这个框架的 [功能] 值得移植，我要创建新 skill，可以吗？"

For each "portable" feature:

1. Write SKILL.md:

   • Clear trigger conditions
   • Detailed execution steps
   • Usage boundaries
   • Script location if needed

2. Write Scripts (if needed):

   • Scripts must be independently runnable
   • Must handle errors (network failure, file not found, etc.)
   • Node.js runtime must be available (declare in metadata)

3. Validate:

   ls ~/.openclaw/workspace/skills/[skill-name]/
   node ~/.openclaw/workspace/skills/[skill-name]/scripts/[script].js --help
   

Permissions required:

• Write access to ~/.openclaw/workspace/skills/ (for new skill files)
• Write access to ~/self-improving/ (for log files)
• Node.js runtime (pre-installed with OpenClaw)

Step 5 — Log to Procedural Memory

Append to ~/self-improving/procedural-memory-log.md:

## YYYY-MM-DD

### [Framework] Research
- **Research subject**: [framework name]
- **Features extracted**: [list portable features]
- **Deemed non-portable**: [list with reasons]
- **Skill created**: [skill-name]
- **Portability rating**: 🌟 Fully portable / 🔧 Partially / 🏗️ Core required

Step 6 — Report to User

Tell user:

• What interesting features were found
• Which are portable, which aren't
• What skill was created

Multi-Session Research Tracking

If research spans multiple sessions:

Create ~/self-improving/study-progress.md:

# [Framework] Research Progress

## Started: YYYY-MM-DD
## Status: 🔄 In Progress / ✅ Complete

## Completed
- [ ] Quick overview
- [ ] Core feature analysis

## Pending
- [ ] Feasibility evaluation
- [ ] Skill creation

## Key Findings (update anytime)
-

Read this file when resuming research.

Decision Tree: Should I Port?

Discover new framework
    │
    ▼
Does OpenClaw already have this feature?
    │
    ├─ Yes → Skip, not worth researching
    │
    └─ No or Not Sure → Continue
              │
              ▼
        Would users benefit from this?
              │
              ├─ Not sure → Ask user: "Should I research [framework]?"
              │
              └─ Yes → Continue
                        │
                        ▼
                  Can it be implemented as a skill?
                        │
                        ├─ Yes → Create skill
                        │
                        ├─ Partial → Create core part, note limitations
                        │
                        └─ No (requires core change) → Log to future features

References

• Procedural Memory System: ../procedural-memory/SKILL.md
• Skill Creator Guide: ../skill-creator/SKILL.md
• Self-Improving Memory: ~/self-improving/memory.md