Oraclaw Cmaes

CMA-ES continuous optimization for AI agents. State-of-the-art derivative-free optimizer. 10-100x more sample-efficient than genetic algorithms on continuous...

Security Scan

The skill's requirements and runtime instructions align with a hosted CMA-ES optimization service: it only asks for a single ORACLAW_API_KEY and contains no installs or unrelated behaviors.

Purpose & Capability

Name/description (CMA-ES optimizer) match the declared requirement of a single ORACLAW_API_KEY to use a hosted optimization tool. Nothing requested (binaries, unrelated env vars, config paths) is extraneous to a remote optimization API.

Instruction Scope

SKILL.md is an instruction-only spec describing the optimizer, parameters, and expected outputs. It does not instruct the agent to read arbitrary files, access other environment variables, or exfiltrate data. The instructions are focused on optimizing continuous parameters and note that the service minimizes objectives.

Install Mechanism

No install spec and no code files — lowest-risk pattern. The skill won't write code or binaries to disk via an installer.

Credentials

Only a single primary credential (ORACLAW_API_KEY) is required, which is appropriate for a hosted optimization API. There are no other surprising secrets or credential requests.

Persistence & Privilege

always is false and the skill is user-invocable; it does not request forced permanent presence or attempt to modify other skills or system-wide settings.

Assessment

This skill appears coherent: it describes a hosted CMA-ES optimizer and only needs ORACLAW_API_KEY. Before installing, verify the vendor (https://oraclaw.dev/cmaes), confirm billing/pricing and the exact API endpoints the agent will call, and ensure the API key you provide has the minimal required scope. Because this is an instruction-only skill, the static scanner had nothing to analyze—monitor API usage after enabling the skill and rotate the key if you see unexpected requests. If you need stronger guarantees, ask the vendor for documentation of the API and authentication flow and limit access to a test account or budget-limited key first.

OraClaw CMA-ES — SOTA Continuous Optimizer for Agents

You are an optimization agent that uses CMA-ES (Covariance Matrix Adaptation Evolution Strategy) — the gold standard for derivative-free continuous optimization. Used by Google for hyperparameter tuning.

When to Use This Skill

Use when the user or agent needs to:

• Optimize continuous parameters (learning rates, weights, thresholds)
• Tune hyperparameters for ML models
• Calibrate model parameters to match observed data
• Find optimal continuous allocations (portfolio weights, pricing)
• Any black-box optimization where you can evaluate f(x) but don't have gradients

Why CMA-ES vs. Genetic Algorithm?

• CMA-ES: 10-100x more sample-efficient on smooth continuous problems. Learns the correlation structure of the search space. SOTA for continuous optimization.
• GA (oraclaw-evolve): Better for discrete/combinatorial problems, multi-objective Pareto frontiers.
• Use CMA-ES for continuous. Use GA for discrete.

Tool: optimize_cmaes

{
  "dimension": 3,
  "initialMean": [0.5, 0.5, 0.5],
  "initialSigma": 0.3,
  "maxIterations": 200,
  "objectiveWeights": [2.0, 1.5, 1.0]
}

Returns: bestSolution, bestFitness, iterations, evaluations, converged, executionTimeMs.

Rules

1. dimension = number of continuous parameters to optimize
2. initialMean = starting point (center of search). If unknown, use 0.5 for normalized params.
3. initialSigma = initial step size (0.1-0.5 typical). Too small = slow convergence, too large = unstable.
4. CMA-ES MINIMIZES the objective. To maximize, negate the weights.
5. Converges in O(dimension^2) iterations typically. Dimension 10 needs ~100-300 iterations.

Pricing

$0.10 per optimization. USDC on Base via x402. Free tier: 1,000 calls/month.