jurisdictional-skill
CLAUDE.md0 starsChanged 17 days agoMITView on GitHub
Start with PLAN.md. A Claude Code plugin for jurisdictional.org containing the gather skill — a conversational interface for exploring local government structure and contributing sourced civic data. The skill is a thin client of the Jurisdictional MCP server (~/workspace/jurisdictional, lib/jurisdictional/mcp/tools/). The server's live tools/list is the source of truth for tool names, fields and enums; when it and this repo disagree, fix this repo.
- Sends data out
In this repositoryOne other file it gives its agents.
- gatherSkill
The file
We split this CLAUDE.md into the rules it gives an agent
4 rules. About 0 are common, 1 is distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 28
Access tiers: no token → 19 public read tools.
- Distinctive: line 36
Keep the three stops (after Step 2, each Step 4 review, the next target) and the Step 4 review table.
- Very specific: line 37
Keep SKILL.md under 500 lines.
- Very specific: line 38
Check enums against the live schemas: curl -s -X POST https://jurisdictional.org/api/mcp/message -H 'content-type: application/json' -d …
≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository. Start with PLAN.md.
What This Is
A Claude Code plugin for jurisdictional.org containing the gather skill — a conversational interface for exploring local government structure and contributing sourced civic data.
The skill is a thin client of the Jurisdictional MCP server (~/workspace/jurisdictional, lib/jurisdictional/mcp/tools/).
The server's live tools/list is the source of truth for tool names, fields and enums; when it and this repo disagree, fix this repo.
Repository Structure
.claude-plugin/plugin.json— plugin manifest (bumpversionon every release).claude-plugin/marketplace.json— single-plugin marketplace:/plugin marketplace add jurisdictional/jurisdictional-skill.mcp.json— the MCP server, withheadersHelper→bin/jurisdictional-headers.shbin/jurisdictional-headers.sh— sendsAuthorization: Bearer <~/.jurisdictional/token>when the file holds ajur_...key,{}otherwiseskills/gather/SKILL.md— the flow: preflight → location → what exists → research → review → submit → verifyskills/gather/references/tools.md— what each MCP tool does to live data, and who can call itskills/gather/references/services.md— gathering servicesskills/gather/schema.md+examples/austin-tx/— local-file format; must validate againstjurisdictional-data/schemas/TESTING.md— scenarios with checks
Key Concepts
- Juricode: given an address, every jurisdiction (federal → local) that governs that location.
- Access tiers: no token → 19 public read tools. Signed in →
gather,propose_change,create_entity,import_jurisdiction, image candidates. Admin →reconcile_roster,end_tenure,merge_entities,delete_entity, coverage reports. - Auth: no header or an empty bearer gets the public tools; an unknown key gets HTTP 401 for the whole connection, so the helper sends no header unless the file holds a
jur_...key. Plugin helpers don't seeTOKEN-like env vars, so the token lives in~/.jurisdictional/token(from/auth/cli). - Writes:
create_entityandimport_jurisdictionapply live;propose_changequeues for review, or corroborates when the value is unchanged. - Provenance: values come only from the government's own page — fetched with
gather, an archived capture (mcc_readfrom the local Common Crawl server in~/workspace/municipal-common-crawl, orgather's archived copy; cited with the capture date), the user's browser, or the user's paste — or from the user's own knowledge. Never search snippets or training data.
When Modifying the Skill
- Invoked via
/gather [address or city, state]; arguments arrive as$ARGUMENTS.disable-model-invocation: truekeeps it user-invoked. - Keep the three stops (after Step 2, each Step 4 review, the next target) and the Step 4 review table; no other pauses.
- Keep SKILL.md under 500 lines; put detail in
references/(one level deep). - Check enums against the live schemas:
curl -s -X POST https://jurisdictional.org/api/mcp/message -H 'content-type: application/json' -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'. - Validate:
claude plugin validate ., and ajv againstjurisdictional-data/schemas/for the examples (see the commit that fixed them).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

