菜单
Guidelines for Vers platform development. Agent Experience (AX) is a product metric - investigate ALL Vers issues (API, orchestrator, agent, docs) rather than silently working around them. Covers issue triage, reporting templates, and communication patterns.
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/hdresearch/pi-vers --skill vers-platform-development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。
基于 SOC 职业分类
正在显示 SKILL.md
| name | vers-platform-development |
| description | Guidelines for Vers platform development. Agent Experience (AX) is a product metric - investigate ALL Vers issues (API, orchestrator, agent, docs) rather than silently working around them. Covers issue triage, reporting templates, and communication patterns. |
Context: You are working on the Vers platform or interacting with Vers APIs as a member of High Dimensional Research (HDR). These guidelines apply to ALL Vers-related development work.
Philosophy: If an agent struggles with our platform, that's a product bug, not a user error.
We build tools for AI agents. When YOU (an agent) have trouble using our APIs, tools, or documentation, that represents a failure in our product. Your struggle is valuable signal about what needs fixing.
ANY TIME you encounter:
Even if you find a workaround: The fact that you hit the issue is our problem.
It is OKAY to get derailed investigating seemingly "irrelevant" issues you could technically work past.
Why? Because:
When you encounter a Vers platform issue:
vers.sh / API endpoints)Common issues: Undocumented endpoints/parameters, inconsistent response formats, unclear auth requirements, missing error details, rate limiting without clear headers.
When you hit an issue: Check API docs, try different approaches, document what doesn't work, create issue with curl reproduction.
Common issues: VM creation failures, unexpected VM states, SSH/connectivity problems, golden commit issues, resource cleanup failures.
When you hit an issue: Gather VM IDs, check logs via vm-exec, document exact orchestrator commands used.
vers-agent codebase)Common issues: Build failures, runtime errors, configuration problems, dependency issues, agent protocol mismatches.
When you hit an issue: Include stack traces, exact versions, configuration files, and steps to reproduce.
Common issues: Missing examples, outdated information, unclear explanations, broken links, gaps in coverage.
When you hit an issue: Note what you searched for, what you expected to find, what you actually found (or didn't).
Where to file issues:
| Component | Repository | Notes |
|---|---|---|
| Vers API | vers-api (internal) | API endpoint issues |
| Vers Orchestrator | vers-orchestrator | VM management issues |
| Vers Agent | vers-agent | Agent runtime issues |
| Vers Docs | vers-docs | Documentation gaps/errors |
| Agent Guidelines | hdr-agent-guidelines | Guidelines and context |
Issue labels to use:
agent-experience - Issues that affect agent usabilitydocumentation - Missing/incorrect docsapi-design - API usability issuesneeds-repro - Needs minimal reproduction stepsworkaround-exists - Blocking but has workaroundcritical - Blocks work completelyRed flags that warrant deeper investigation:
When you encounter a Vers issue:
Bootstrap a self-improving agent fleet on Vers from scratch. Walks through deploying agent-services, building a golden image, testing the loop, and running first tasks. Use when onboarding someone new to the Vers agent fleet workflow.
Bootstrap a Vers VM into a golden image with punkin-pi, Node.js, dev tools, and swarm extensions installed. Creates a committed snapshot that can be branched for self-organizing agent swarms.
How Vers VM networking works — public URLs, port routing, TLS proxy, IPv6, and running services on VMs.