| name | Self-Evolving Systems |
| description | Automated harness evolution using MetaHarness propose-evaluate-log-iterate loops, Karpathy autoresearch patterns, and the Bitter Lesson applied to agentic orchestration |
Self-Evolving Systems
The harness around the model matters as much as the model itself. A 6x performance gap exists between the best and worst harness for the same weights. Hand-engineering harnesses is the status quo. It shouldn't be.
Core Capability
Design, run, and audit self-improvement loops for any agentic system component — workflows, prompts, retrieval logic, memory management, orchestration patterns. Uses the MetaHarness methodology: a coding-agent proposer that inspects prior code + execution traces, diagnoses failure modes, proposes targeted edits or full rewrites, and iterates until the system converges on Pareto-optimal performance.
This is not self-correction (recovering from errors). This is self-evolution (permanently improving the system).
What This Skill Produces
| Workflow | Output |
|---|
| Self-Evolve | Full MetaHarness-style evolution loop on any Antigravity component |
| Harness Evolve | Focused evolution of a specific workflow's prompts, flow, and gates |
| Auto-Experiment | Karpathy-style experiment runner with git-branch isolation |
| Evolution Audit | History inspection, Pareto frontier, regression detection |
| Skill Anneal | Self-annealing applied to skill prompts/workflows via past failures |
| Proposer Sprint | Time-boxed, fixed-iteration improvement sprint |
| Evolution Status | Dashboard of all active/completed evolution loops |
| Bitter Lesson Check | Audit hand-coded heuristics for evolution replacement potential |
| Trajectory Ratchet | Multi-cycle evolution protocol with monotonic progress enforcement — prevents drift, local optima, and coherence loss across hundreds of changes |
When to Deploy
- A workflow/skill has been manually tuned but performance plateaus
- Quality gate scores consistently land in the 6-7 range (good but not great)
- The same class of error keeps recurring despite self-annealing fixes
- You want to discover better approaches than hand-tuned prompts
- Before upgrading to a more expensive model (evolve the harness first)
- Periodic optimization sprints on high-value workflows
Stacking Opportunities
- + Nick Saraev (Agentic Workflows): Self-annealing + evolution = self-correcting AND self-improving
- + Nate B. Jones (Orchestration Intelligence): Harness audit identifies gaps → evolution fills them
- + Quality Gate (Chain Step 6): Quality gate failures become the search set for evolution
Key Principles
- Harness before intelligence — Evolve the harness before upgrading the model
- Trace over score — Rich execution traces > scalar reward signals
- Adaptive over monolithic — Let the proposer select what to inspect, don't pre-pack
- Code-space regularity — Represent harnesses as programs; coding models avoid brittle solutions
- Skill text is strongest lever — The quality of the skill/workflow matters more than iteration count
- The Bitter Lesson — Hand-coded heuristics always lose to learned solutions eventually
Load Order
- Load this
SKILL.md (required — methodology + principles)
- Load
genius.md for deep patterns, hidden knowledge, and the full MetaHarness methodology
- Load specific workflow for task
- Cross-reference
directives/deep_self_annealing.md for Tier 4 evolution triggers
