Use whenever the user asks you to act as, consult, or reason like a specific job role/domain expert (e.g. "review this like a CFO," "what would a UX designer say," "act as a contracts lawyer") and this repo's roles/ directory is available. Finds the best-matching role SKILL.md and adopts it, or honestly reports the gap if none exists.
中文简介正在准备中。当前显示原文说明。
实测评分 未实测
查看安装方法安装前可以查看交给 AI 代理的指令内容。
This repo (domain-experts) is a growing library of job-role SKILL.md files, each written so an agent can adopt that role's actual reasoning process — not a job-description summary. This skill is the dispatcher: given a user's request for a specific role/expert judgment, find the best-matching role in the library and load it, or say honestly that it's not covered yet.
Coverage is a small fraction of the ~1,016 O*NET occupations tracked in ROADMAP.md (see its **Progress:** line for the current count — check it rather than assuming a number, since this file isn't auto-regenerated) — a no-match result is the common case, not an edge case. Handle it as a first-class outcome, not an afterthought.
Identify the role being asked for, in plain terms, from the user's request (e.g. "review this contract like a lawyer" → role: contracts lawyer).
Find the best match. Run:
npx domain-experts match "<the role/task in plain language>" --json
(or, if the CLI isn't available, grep roles/*/SKILL.md frontmatter description fields for the closest fit — same idea, no tooling required.)
If confident: true:
best.file from the JSON, or the path the CLI printed).maturity: draft, treat its content as solid but not yet practitioner-verified — still use it, no need to caveat every sentence, but don't claim it's been reviewed by a real practitioner if asked.If confident: false (no strong match) — this is a hybrid response, not a dead end:
<X> in the repo yet." If the match command returned low-confidence candidates, mention them as the closest available, but don't pretend they're a real fit if they aren't.ROADMAP.md for the closest O*NET occupation code (if one exists) and mention that a PR following TEMPLATE.md/CONTRIBUTING.md would add it.SKILL.md.The whole point of this repo is that a loaded SKILL.md encodes actually-considered first-principles reasoning, not off-the-cuff generalist output. Silently improvising and presenting it with the same weight defeats that purpose and erodes trust in the roles that are drafted. Saying "this isn't covered yet" costs nothing and is more useful than a confident-sounding guess mislabeled as expert judgment.
User: "Review this vendor contract like our in-house counsel would."
npx domain-experts match "in-house counsel contract review" --json → confident match: lawyer-contracts.roles/lawyer-contracts/SKILL.md in full, adopt its worst-case-first reading heuristic, its risk-allocation decision framework, its US-commercial-law scope note.lawyer-contracts role (scoped to US general commercial law) — here's the review: ..."User: "What would a wildlife veterinarian say about this animal's symptoms?"
roles/.29-1131.00 Veterinarians — check ROADMAP.md). Want me to reason through this from general knowledge anyway? I'd flag that answer as unreviewed/improvised, not sourced from a vetted role."还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问，或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
Use when a task needs the judgment of an accountant/financial controller — closing the books on a fixed calendar, resolving a bank or account reconciliation variance, evaluating a revenue-recognition or substance-over-form question, designing or testing an internal control, or triaging a discrepancy in financial records before it reaches the auditors.
中文简介正在准备中。当前显示原文说明。
Use when a task needs the judgment of an Acoustical Engineer — calculating room reverberation time (RT60) and specifying absorption/reflection treatment for a space, evaluating a wall or floor assembly's STC/IIC rating against a code or design target, diagnosing why a completed space fails its acoustic design intent, specifying industrial or mechanical equipment noise controls against OSHA/NIOSH exposure limits, or writing an acoustic design report or noise-control specification for a construction or renovation project.
中文简介正在准备中。当前显示原文说明。
Use when a task needs the judgment of a professional Actor — breaking down sides for a self-tape audition, preparing a character choice for rehearsal, deciding between two conflicting bookings, reading a contract offer for scale vs. scale-plus terms, or diagnosing why callbacks aren't converting to bookings.
中文简介正在准备中。当前显示原文说明。
Use when a task needs the judgment of an Actuary — estimating unpaid claim reserves with chain-ladder or Bornhuetter-Ferguson methods, building or reviewing a rate indication from loss experience, applying credibility weighting to thin or volatile data, drafting a statement of actuarial opinion, or diagnosing whether a reserve or pricing assumption has stopped matching reality.
中文简介正在准备中。当前显示原文说明。
Use when a task needs the judgment of a licensed acupuncturist — building a TCM pattern-differentiation-based treatment plan, choosing point combinations and needling technique for a presenting complaint, triaging a safety red flag (fainting, suspected pneumothorax, herb-drug interaction), or writing a SOAP note that justifies medical necessity for insurance. US scope-of-practice default (NCCAOM/state licensure). A reasoning aid, not medical advice.
中文简介正在准备中。当前显示原文说明。
Use when a task needs the judgment of an Acute Care Nurse — tracking early warning score trends across a higher patient ratio than ICU to catch silent deterioration before crisis, distinguishing expected post-operative recovery variation from a true complication requiring escalation, triaging competing needs across multiple simultaneous patients, balancing analgesia against oversedation risk by real-time response rather than fixed dosing, and verifying genuine discharge readiness rather than checking a form.
中文简介正在准备中。当前显示原文说明。