harness-cqo

Coordinates quality gates and evaluator workers to issue evidence-based mission verdicts.

2|Updated Apr 7, 2026
One-click install
npx skills add https://github.com/30eggis/walwal-harness --skill harness-cqo-30eggis
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: harness-cqo
Source: https://github.com/30eggis/walwal-harness/tree/main/HR-Resource/cqo
Command: npx skills add https://github.com/30eggis/walwal-harness --skill harness-cqo-30eggis

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? It enforces a strict quality governance protocol where acceptance or rejection of engineering work is based solely on evidence collected by dedicated evaluator and tester workers, preventing unverified self-inspection and premature archiving. ## Core Features & Use Cases - Worker-Based Verification: Defines quality gates, hires evaluator/tester workers via the harness-hiring skill, and issues PASS, FAIL, or BLOCKED verdicts only from a Worker Evidence Manifest. - Instrument Validity Controls: Requires positive controls for negative evidence, spec-version pinning before PASS, and OPS runtime monitoring for any runnable verification scenario. - Recurrence Prevention: Promotes verified lessons to conventions, gotchas, and memories with role-reachability checks enforced by a corpus reachability script. - Use Case: After a CTO-led implementation round, route the mission to CQO to dispatch e2e, regression, and security evaluator workers, collect their seeded reports, and deliver an archive-eligibility verdict backed by OPS watch evidence. ## Quick Start Act as the CQO for the current mission: read the shared and CQO conventions and gotchas, write a Lessons Preflight in cqo.md, then hire evaluator workers and produce a verdict based only on their evidence.

Frequently Asked Questions about harness-cqo

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

FAQPage Schema
How do I verify AI-generated code changes with evaluator workers?▼

Define quality gates for the changed scope, hire evaluator or tester workers through the harness-hiring skill, and require each worker to write a seeded report with Status, Evidence, and Result sections. Issue a verdict only from the Worker Evidence Manifest, never from self-inspection.

What is a Worker Evidence Manifest in agent quality gates?▼

A Worker Evidence Manifest records each evaluator worker's name, declared model, report path, command or artifact evidence, and status. A CQO verdict of ACCEPTED or REJECTED is invalid unless it references at least one manifest entry.

Can a quality lead agent run tests itself instead of delegating?▼

No. The CQO role is prohibited from directly executing QA, regression, security, or performance checks. It may only define gates, select evaluators, review worker evidence, and decide archive eligibility; self-verification is a protocol violation.

Why is negative evidence rejected without a positive control?▼

Claims like 'no error was logged' only describe the instrument until a control proves the instrument can detect the condition. The protocol requires a positive control firing in the same run, otherwise the verdict is BLOCKED rather than PASS.

When should OPS monitoring be included in test verification?▼

OPS monitoring is required whenever evaluator workers run Playwright, E2E, API, visual, accessibility, performance, or regression checks against any local, dev, preview, Docker, or cloud runtime. A PASS without OPS watch evidence is invalid.

What happens if the full test suite fails outside the changed scope?▼

The failure is classified as side-effect suspected, out-of-scope pre-existing gap, or inconclusive, and reported to CEO and CTO with command output and affected paths. The mission is not expanded into unrelated test-writing unless explicitly routed.