菜单
Use whenever the user asks about a motor-vehicle statute, citation, contributing factor, OR a Canadian personal-injury fact pattern — always query Specter's Harvester API before answering. The API auto-routes between two collections (US statutes + Canadian PI case law).
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/ThomasMoreAI/legal-skills-open --skill harvester-query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。
Use when retrieving arbitration institutional rules (ICC, LCIA, SCC, SIAC, HKIAC, VIAC, МКАС/МАК при ТПП України, UNCITRAL) — fetching current version, verifying redaction applicable to the date of arbitration agreement, constructing URLs for official rule texts
Use when choosing the Polish legal regime for letters, requests, applications, complaints, petitions, public-information requests, KPA filings, PPSA complaints, RODO access requests, registry extracts, court-file access, tax/ZUS/cudzoziemcy/USC procedures, or professional lawyer letters. Prevents mixing UDIP, KPA, PPSA, RODO, registry, special-procedure, and advocate/radca letter regimes.
Use when preparing applications for recognition and enforcement of foreign arbitral awards in Poland, applications for setting aside arbitral awards under KPC art. 1205–1211, or opposing such applications — mapping Article V of the 1958 New York Convention to art. 1214–1215 of the Polish KPC, identifying grounds for refusal, structuring public policy arguments
基于 SOC 职业分类
正在显示 SKILL.md
| name | harvester-query |
| title | harvester-query |
| description | Use whenever the user asks about a motor-vehicle statute, citation, contributing factor, OR a Canadian personal-injury fact pattern — always query Specter's Harvester API before answering. The API auto-routes between two collections (US statutes + Canadian PI case law). |
| author | barndonmai |
| author_url | https://github.com/barndonmai/specter/tree/main/openclaw/skills/harvester_query |
| license | MIT |
| version | 0.1.0 |
| execution_mode | open |
| jurisdiction | cross-jurisdiction |
| practice | personal-injury |
| language | en |
Specter's Harvester is a FastAPI service that fronts two Chroma collections:
| Collection | Contents | When the API uses it |
|---|---|---|
specter_statutes | 384 US motor-vehicle statutes (CA, TX, FL, PA, IL, OH) | default — any US-jurisdiction or generic query |
canada_pi_cases | 340 real Canadian PI cases (CanLII, with damages, injury type, plaintiff_won, deciding_factor) | auto-routed when the query mentions Canada / Ontario / Quebec / Toronto / ONSC / ONCA / canlii / etc., or when ?jurisdiction=Canada or ?state=CA-CAN is passed |
Every citation the user sees should come from the Harvester (or, on miss, from a verified web source). Never fabricate.
search first; ask follow-ups only after showing what hit.lookup_citationsearchaskfactor filter (see categories below).For Canadian fact-pattern queries especially, call first with whatever the user said, even if it's vague. Show the top hits, THEN ask any follow-ups. The Canada corpus is fact-pattern-indexed — thin queries still return relevant cases. Demanding more facts before searching wastes the corpus.
searchThe helper module sits next to this SKILL.md:
import sys
from pathlib import Path
sys.path.insert(0, str(Path(__file__).parent)) # if running ad hoc
from client import lookup_citation, search, ask, healthz, list_factors
In OpenClaw, prefer running it via exec:
cd ~/Specter && python3 -m openclaw.skills.harvester_query.client lookup "Cal. Veh. Code § 23152(a)"
cd ~/Specter && python3 -m openclaw.skills.harvester_query.client search "drunk driving" --state CA --factor "DUI/DWI"
cd ~/Specter && python3 -m openclaw.skills.harvester_query.client ask "what statutes cover hit and run in CA"
Or from Python:
from openclaw.skills.harvester_query.client import lookup_citation, search, ask
lookup_citation(citation: str) -> dict | NoneExact match by citation string. Returns the full record or None if not in the Harvester.
search(q: str, *, state=None, factor=None, pi_only=False, k=10) -> list[dict]Semantic search. state is a 2-letter code ("CA", "TX"). factor must be one of the 17 canonical categories.
ask(question: str, *, state=None, factor=None, k=8) -> list[dict]Same as search but framed as a question. Use this when the user hands you a natural-language query.
healthz() -> dictCheap liveness check. Use to confirm the API is running before reporting a miss.
list_factors() -> list[str]The 17 categories. Useful for synonym mapping in voice.
Specter's voice rules (see SOUL.md): 2–3 lines max by default. Cite, don't quote.
Good:
Cal. Veh. Code § 22350. Basic speed law — drive no faster than reasonable for conditions. leginfo.legislature.ca.gov
Bad:
Of course! Pursuant to California Vehicle Code Section 22350, "no person shall drive a vehicle…" (wall of text)
Always include the source_url from the record. Never include a citation without one — per hackathon ground rules, records without a real source URL don't count.
If lookup_citation returns None or search/ask returns []:
healthz() — if the API is down, say so plainly.Never paper over a miss with a guess.
DUI/DWI
Driving Too Fast For Conditions
Failure to Maintain Lane
Failure to Obey Traffic Control Device
Failure to Use/Activate Horn
Failure to Yield at a Yield Sign
Failure to Yield the Right-of-Way
Fleeing a Police Officer
Fleeing the Scene of a Collision
Following Too Closely
Improper Lane of Travel
Improper Passing
Improper Starting
Improper Stopping
Improper Turning
Reckless Driving
Using a Wireless Telephone/Texting While Driving
If the user uses a synonym, map internally and answer in their words.