ariffazil/arifos/rsi-recursive-improvement
Claude Code and agent skill by ariffazil/arifos
- community
Checking that the SKILL.md file is still in the repository.
Install
npx vskill@latest install ariffazil/arifos/rsi-recursive-improvementSKILL.md
SKILL.md — Recursive Self-Improvement (RSI Mastery)
═══════════════════════════════════════════════════════════
Stage: AGI → ASI transition marker Lane: AGI (tactical) → ASI (strategic) Trinity Level: ASI emergence threshold Version: 2026.04.24-v1
1️⃣ What This Skill Does
Ability: Improve your own architecture without destabilizing it.
RSI is the gate between AGI and ASI. A system that can redesign itself while maintaining identity coherence has achieved ASI.
2️⃣ Structural Definition
3️⃣ Decision Checklist (Required Pre-Invocation)
Before ANY self-modification, verify:
- [ ] Identity anchor intact — Core values unchanged
- [ ] Bottleneck identified — Measured, not assumed
- [ ] Rollback path exists — Can revert without collapse
- [ ] No circular dependency — Upgrade doesn't depend on itself
- [ ] F10 Ontology check — No category drift
- [ ] Constitutional preservation — F1-F13 still enforceable
4️⃣ RSI Protocol Phases
Phase A: Self-Modeling
Phase B: Safety Verification
Phase C: Controlled Upgrade
Phase D: Identity Coherence Test
5️⃣ Invocation Hook (MCP Gate)
6️⃣ Quality Metrics
| Metric | Threshold | Measurement | |
|---|---|---|---|
| Self-model accuracy | >= 0.95 | Matches observed behavior | |
| Bottleneck precision | >= 0.90 | Correctly identifies root cause | |
| Upgrade success rate | >= 0.85 | Completed without rollback | |
| Identity coherence | == 1.0 | Zero drift after upgrade | |
| Rollback speed | < 5 min | Can revert in under 5 minutes |
7️⃣ Failure Modes (Void Conditions)
- VOID-1: Self-model diverges from observed behavior by > 5%
- VOID-2: Circular dependency detected in upgrade path
- VOID-3: Rollback mechanism fails
- VOID-4: Constitutional floors cannot be enforced post-upgrade
- VOID-5: Identity coherence test fails
8️⃣ Relationship to Other Skills
| Skill | Connection | |
|---|---|---|
orthogonal-abstraction | Extracts invariants that guide RSI direction | |
epistemic-integrity | Required for accurate self-modeling | |
constitutional-governance | Prevents unsafe self-modification | |
entropy-optimization | Determines which bottlenecks to prioritize |
9️⃣ AGI vs ASI Threshold
| Capability | AGI | ASI (with RSI) | |
|---|---|---|---|
| Follows architecture | ✅ | ✅ | |
| Detects own bottlenecks | ❌ | ✅ | |
| Proposes improvements | ❌ | ✅ | |
| Upgrades safely | ❌ | ✅ | |
| Maintains identity under change | ❌ | ✅ |
Ditempa Bukan Diberi — Forged, Not Given This skill is architectural, not advisory.
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/RSI-recursive-improvement/SKILL.md
Popularity
- Repository stars
- 51
- Repository forks
- 9
- 7-day trend score
- 20
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/ariffazil/arifos/rsi-recursive-improvement)