skills.rest
AI-Lab-for-Cities-at-Harvard
AI-Lab-for-Cities-at-HarvardOFFICIAL
Explore all skills by this author →
GitHub

jurisdiction-priority-scan

Scans open data, media, and testimony to rank agency priority problems.

2|2|Updated Aug 29, 2026
Search Tags:#performance management#government#opencontext#open data#problem identification#priority scan#civic analytics
One-click install
npx skills add https://github.com/AI-Lab-for-Cities-at-Harvard/civic-skill-exchange --skill jurisdiction-priority-scan-ai-lab-for-cities-at-harvard
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: jurisdiction-priority-scan
Source: https://github.com/AI-Lab-for-Cities-at-Harvard/civic-skill-exchange/tree/main/skills/sgarcese/jurisdiction-priority-scan
Command: npx skills add https://github.com/AI-Lab-for-Cities-at-Harvard/civic-skill-exchange --skill jurisdiction-priority-scan-ai-lab-for-cities-at-harvard

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Performance officers and agency leaders need a defensible, evidence-backed list of problems to prioritize, but manually scanning open data portals, local news, and public testimony is slow and prone to unverified claims. This Skill runs a broad, disciplined scan and produces a classified, confidence-flagged candidate list. ## Core Features & Use Cases - Multi-source retrieval: Queries the jurisdiction's open data portal through an OpenContext MCP connector, plus local media, public testimony, audits, and social media where a tool can reach it. - Structured classification: Tags every candidate by evidence source, shape (acute/chronic), horizon, degree of control, and dominant cost, with a Verified/Partial/Unverified confidence flag. - Two-list output: Produces List A (core problems inside the agency's remit) and List B (adjacent problems where the agency has leverage), plus a memo with an analyst work plan. - Use Case: A city performance office asks what the 311 operation should focus on next quarter; the Skill queries service request data, searches local coverage and council testimony, and returns a ranked table with citations and verification gaps. ## Quick Start Run a priority scan for the city's 311 agency using the configured OpenContext connector and produce the classified list and memo.

Frequently Asked Questions about jurisdiction-priority-scan

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a priority scan for a city or state agency?▼

Configure the skill with a domain pack, the principal agency's remit, and an OpenContext connector, then ask for a priority scan. It retrieves performance data, local media, and testimony in parallel, verifies figures, and outputs a classified list and memo.

What data sources does the jurisdiction priority scan use?▼

It uses four sources: the jurisdiction's open data portal via an OpenContext MCP connector, local media coverage, public testimony from hearings and budget sessions, and social media where a tool can reach it. Audits and federal data support as Web sources.

Which open data portals and backends are supported?▼

Any portal exposed through an OpenContext MCP connector works, covering Socrata, CKAN, ArcGIS Hub, and Opendatasoft backends. Each backend has different query verbs, dialects, and row limits documented in references/backends.md.

Can I use this scan without an open data connector?▼

It is not recommended. Without a connector the scan degrades to media and testimony alone, which over-weights whatever received attention, and the skill's own guidance says to avoid that scenario.

Why do scan results include Verified, Partial, and Unverified flags?▼

The flags enforce the rule that no number appears without a retrieval in the current run. Verified figures came from a tool call, Partial items have adjacent evidence, and Unverified items are judgment with a stated check to confirm them.

When should I use the quick-scan variant instead?▼

Use a quick-scan variant for live demos or workshops on a three-to-four-minute clock. The full scan takes seven to ten minutes and is meant for desk research and preparation.

Related Skills

Explore More Data & Analytics Skills→
akillnessakillness
40

opencontext

Store and retrieve persistent context across sessions and repositories.

Community
Advanced
smanistsmanist

orient-simple

Select the initial task by reading PROJECTS/*/TASKS.md and checking git status.

Community
Intermediate
miladdavoodi9miladdavoodi9

trade-compare

Compare two stocks across valuation, growth, profitability, risk, and sentiment dimensions.

Community
Advanced
Umbral-BotUmbral-Bot

linkedin-content

Generate LinkedIn post templates and strategies for BIM/AEC consultants.

Official
Intermediate