菜单
Pi-native evaluate→improve→persist harness. Provides a systematic loop for generating candidates, judging them against rubrics, and accumulating knowledge as playbooks. Use for iterative output improvement, not for one-off edits (use continuity) or single PR review (use zereight-review).
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/zereight/skill --skill harness命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Audit a UI or design against WCAG 2.2 AA/AAA and ARIA patterns, returning criterion-referenced findings with severity and specific fixes. Use when the user wants an accessibility check, contrast verification, keyboard/screen-reader review, or wants to confirm a component meets POUR.
Apply a visual direction — an archetype (high-end agency, editorial minimal, brutalist, soft-SaaS, dark-tech) or one of 138 named design systems (apple, linear-app, stripe, vercel, notion, material, shadcn, spotify, tesla…) — by resolving it into the token system. Use when the user wants a specific look/vibe/brand feel, or asks to make a design feel premium/expensive/non-generic.
Create and maintain BankX deep link request tickets from PRDs, specs, notification requirements, or screen requests. Use when the user asks to request BankX deeplinks, create MAR Jira tickets, extract deeplink needs from PRDs/docx files, follow the "How to request deep link" Confluence workflow, or update deep link request issues on bankx.atlassian.net.
基于 SOC 职业分类
正在显示 SKILL.md
| name | harness |
| description | Pi-native evaluate→improve→persist harness. Provides a systematic loop for generating candidates, judging them against rubrics, and accumulating knowledge as playbooks. Use for iterative output improvement, not for one-off edits (use continuity) or single PR review (use zereight-review). |
| argument-hint | [solve|knowledge|status] |
Builds on autoctx (judge/improve core) + Pi tools to provide a coherent harness.
| Layer | Source | Role |
|---|---|---|
| Extension | ~/.pi/agent/extensions/harness/index.ts | Tools + TUI + lifecycle |
| Judge engine | autoctx (via pi-autocontext dep) | LLM-based evaluation |
| Skill | .agents/skills/harness/SKILL.md | Usage guide + routing |
| Tool | Description |
|---|---|
harness_solve | Full loop: generate → judge → improve → persist playbook |
harness_knowledge | Read/write/list playbooks (knowledge/<scenario>/playbook.md) |
harness_status | Recent runs, scores, and scenarios overview |
# Solve a task with rubric-driven improvement
goal: "Improve zereight-review checklist for RN PRs"
rubric: "Actionable findings, Effect anti-patterns, security when relevant"
→ Best output + playbook saved to knowledge/<scenario>/
# Check previous runs
→ Recent runs with scores and scenarios
# Browse accumulated knowledge
action: list → all scenarios with playbooks
action: read, scenario: zereight-review → full playbook
| Need | Use |
|---|---|
| Systematic output improvement | harness_solve |
| Review accumulated lessons | harness_knowledge read |
| Check past run history | harness_status |
| One-off judge/improve | autocontext_judge / autocontext_improve (pi-autocontext) |
harness_solve consumes Pi provider quota per generation round.gens: 1 to calibrate rubric, then increase.harness_knowledge and harness_status are zero-cost.