菜单
Evaluate skills — trigger accuracy, fidelity, and value. Use when asked to test skills, run evals, or check if skills work.
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。
基于 SOC 职业分类
Business, outreach and public-relations manager for AI projects and technical teams. Use when an agent must turn a collaboration opportunity, partner request, portfolio update, public claim, outreach message, follow-up, sponsorship path or service proposal into a bounded decision, safe communication package and tracked next action without sending or publishing autonomously.
Deploy paper content from configured site_ready files to d-nd.com VPS. Handles md-to-HTML conversion, bilingual merge, figure path resolution, and pages.json update.
Five mechanical gates for any content publish pipeline with CMS + rendering layers. Prevents false security: 'API returned 200' does not mean 'visitor sees clean content'. Use when writing content to a multi-layer serving system (CMS API, static files, prerendered HTML, cached copies).
Multi-node consultation protocol for high-leverage decisions. Dispatches the same question to N independent LLM/agent nodes in isolation, then synthesizes their responses into a summa that exposes convergence (high-confidence claims), dissensus (real uncertainty zones), and emergent points (insights no single node produced). Reduces single-node training bias. Supports recursive escalation for stable-state convergence. Use for decisions that propagate via A14 cascade — seed updates, crystallizations, advisory→mechanical promotions, high-visibility copy, lab result interpretation.
Closure reflection protocol. After a significant work block concludes (feature shipped, session ending, major commit landed, cross-node coordination resolved), runs a 10-question interview that extracts meaning, impact, and next questions — then emits multiple audience-specific artifacts (changelog, external editorial, AI integration docs, memory crystal, backlog seed). Turns implicit maturation into explicit narrative. Use at the end of meaningful work, not after trivial edits.
Pre-commit check for public-facing copy (knowledge base definitions, page content, docs). Detects apologetic hedging — phrases that declare 'degrees of truth' (possible/necessary, current/future, one-of-many/the) and open a dualistic framing the model transcends. Use when drafting or reviewing any copy that describes the model, its transductions, or its tools.
| name | eval |
| description | Evaluate skills — trigger accuracy, fidelity, and value. Use when asked to test skills, run evals, or check if skills work. |
| user-invocable | true |
Every skill has a section ## Eval with:
To test: read every SKILL.md, find ## Eval, verify each test case. Output per skill: Trigger X/Y, Fidelity X/Y, Issues found.
