zenigame-calibrate-gate

Adjust stage_a_gate_stage_b_ratio from gate_stats and persist to config/alpha_factory/default.yaml.

Updated Apr 17, 2026
One-click install
npx skills add https://github.com/kent013/zenigame-fx --skill zenigame-calibrate-gate
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: zenigame-calibrate-gate
Source: https://github.com/kent013/zenigame-fx/tree/main/.claude/skills/_archived/zenigame-calibrate-gate
Command: npx skills add https://github.com/kent013/zenigame-fx --skill zenigame-calibrate-gate

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Stage A Gate calibration reads the previous Run's gate_stats to automatically adjust the stage_a_gate_stage_b_ratio in config/alpha_factory/default.yaml for the next run, using Claude (LLM) to decide on a rational value.

Core Features & Use Cases

  • Reads the latest summary JSON to extract gate_stats such as a_pass_rate, total_gate_blocked, total_a_pass, total_evaluated, and stage_b_ratio.
  • Computes a new stage_b_ratio within safe bounds, applying a ±0.05 change limit and clamping to [0.10,0.30], then updates config/alpha_factory/default.yaml.
  • Generates a transparent report explaining the rationale and the before/after values for governance and reproducibility.

Quick Start

Run the calibration to read the latest summary.json, compute the new stage_b_ratio, and update config/alpha_factory/default.yaml.

Frequently Asked Questions about zenigame-calibrate-gate

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate gate ratio tuning from prior run stats?

Automated stage_b_ratio calibration reads prior run gate_stats like a_pass_rate and total_gate_blocked from a summary JSON, then computes and updates a new ratio in config/alpha_factory/default.yaml for the next run.

What is stage_a_gate_stage_b_ratio calibration in evolutionary algorithm pipelines?

Stage_a_gate_stage_b_ratio calibration adjusts the evaluation gate threshold by analyzing previous run statistics like a_pass_rate and total_evaluated to optimize candidate flow between pipeline stages.

What are the limits when adjusting stage_b_ratio in config/alpha_factory/default.yaml?

Stage_b_ratio adjustments are clamped to a [0.10, 0.30] range and limited to a ±0.05 maximum change from the previous ratio to ensure stable evolutionary algorithm runs.

How do I update stage_b_ratio based on a_pass_rate and total_gate_blocked?

Updating stage_b_ratio involves reading a_pass_rate, total_gate_blocked, total_a_pass, and total_evaluated from the latest summary JSON, then applying the new value with a justification to the default.yaml configuration.

Can I use LLM to decide gate ratio calibration for my automation pipeline?

Yes, Claude (LLM) decides a rational stage_b_ratio value by analyzing previous run gate_stats, generating a transparent report with rationale and before/after values for governance and reproducibility.