harness-creator-adapter

Creates and revises Gemini runtime harness assets with validation and claim boundaries.

Updated May 22, 2026
One-click install
npx skills add https://github.com/handaeho/Harness_Engineering --skill harness-creator-adapter-handaeho
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: harness-creator-adapter
Source: https://github.com/handaeho/Harness_Engineering/tree/main/prompt-stack/v36/gemini/skills/harness-creator-adapter
Command: npx skills add https://github.com/handaeho/Harness_Engineering --skill harness-creator-adapter-handaeho

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Maintaining Gemini runtime harness assets (instructions, state, verification, scope, lifecycle, validation, benchmarks) is error-prone: changes can blur owner layers, overstate verification claims, or mirror autonomous/Codex assets into the wrong location. This Skill provides a disciplined procedure for creating or adapting those assets while enforcing claim boundaries and static validation. ## Core Features & Use Cases - Structured Change Procedure: Follows an Analyze -> Plan -> Create/Revise -> Test -> Verify -> Report workflow for every harness asset change. - Owner-Layer Routing: Routes changes to the correct layer (Gemini runtime, shared state, verification, lifecycle, evidence, archive, or report) using a detailed reference guide. - Claim Boundary Enforcement: Prevents unsupported claims such as provider_verified, release_gated, or production_ready unless separately executed provider evidence exists. - Use Case: When adding a new Gemini CLI layout validation fixture, activate this Skill to place the asset under the correct owner layer, run static validation, and report the claim boundary and remaining provider risk. ## Quick Start Use the harness-creator-adapter skill to add a Gemini runtime validation fixture for the CLI layout and report the validation result and claim boundary.

Frequently Asked Questions about harness-creator-adapter

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

FAQPage Schema
How do I create or modify Gemini runtime harness assets?▼

Activate this Skill and follow the Analyze, Plan, Create/Revise, Test, Verify, Report procedure. Name the harness gap and subsystem, confirm the owner layer, make the smallest artifact change, run static validation, and report the claim boundary.

When should I use the harness-creator-adapter skill?▼

Use it when changing Gemini runtime instructions, state, verification, scope, lifecycle, handoff, validation, benchmark, clean-state, artifact-map, or source-ledger assets. Do not use it for ordinary code fixes, release decisions, pure research, or architecture design.

Can Gemini runtime assets be placed under autonomous or codex directories?▼

No. Gemini runtime assets must live under the gemini/ directory, with CLI deployment using .gemini/GEMINI.md and .gemini/skills. Placing them under autonomous/99_total or codex/ violates the owner-layer rules.

What claims are allowed after changing harness assets?▼

Changes can support local_static_runtime_validation only after the relevant static runner passes. They cannot support provider_verified, adapter_checked, release_gated, production_ready, or live canary claims without separately executed provider evidence.

Does this skill run live provider validation for Gemini API changes?▼

No. The Skill keeps static validation separate from live provider execution. API-specific assertions must be grounded in official Gemini docs via source ledgers, and live provider verification requires separate approved execution.