Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add 2016judea/small-business-intelligence-mcp --skill find-the-public-recordgit clone --depth 1 https://github.com/2016judea/small-business-intelligence-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/2016judea/small-business-intelligence-mcp/find-the-public-record)<a href="https://agentmods.dev/skills/2016judea/small-business-intelligence-mcp/find-the-public-record"><img src="https://agentmods.dev/badge/skills/2016judea/small-business-intelligence-mcp/find-the-public-record/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/2016judea/small-business-intelligence-mcp/find-the-public-record"><img src="https://agentmods.dev/badge/skills/2016judea/small-business-intelligence-mcp/find-the-public-record.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00057 | $0.00627 |
| Opus 5.5 | $0.00023 | $0.00251 |
| Sonnet 5.5 | $0.00011 | $0.00125 |
| Haiku 4.5 | $0.00006 | $0.00063 |
Grade A, and why
find-the-public-record scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 54 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Find the public record
data_source_atlas returns a source-first research plan: which public record
answers the question, how to reach it directly, what the answer is worth, and
what the record cannot answer at all. It does not return the answer. You go and
get that with web search.
The difference this makes: reading whatever a search engine surfaced versus pulling the administrative record that settles it.
Quick start
- Call
data_source_atlaswith the real question, including the city and state. It returns aresearch_procedure, anoutput_schemafor the finished deliverable, aquality_rubricandcaveats. - Execute the
research_procedureyourself with web search. Go to the named record — the county assessor's parcel search, the recorder of deeds, the city's licence or inspection portal, Census CBP or ACS, a BLS series, the state registry — not to an aggregator that resells it. - Produce the deliverable in the shape
output_schemadescribes, and check it againstquality_rubricbefore answering. - State the
caveats. What the record cannot answer is part of the answer.
Rules
Name the record, cite the pull. Every figure carries where it came from and when it was pulled. "The county assessor shows $X, pulled today" beats "$X".
Non-disclosure states have no sale price. In Kansas, Texas, Utah and the other non-disclosure states, the recorded deed does not carry a price. Say so and pivot to what does exist — assessed value, the assessor's own ratio study, mortgage amount — instead of producing an estimate that reads like a record.
Distinguish the record from the vendor. A figure from CoStar, Zillow or a data reseller is a product, not a public record. It can be cited as such but it is not the administrative answer, and its methodology is usually undisclosed.
When the record does not exist, say that. The most valuable output of this tool is often "no public record answers this, and here is why." Commercial rent in Minnesota is the standing example: leases are not recorded anywhere.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 54 lines · 57 tokens per session scan A 50bb0bd807ab
find-the-public-record is a skill published in the GitHub repository 2016judea/small-business-intelligence-mcp (0 stars, last pushed 4d ago), licensed MIT. It adds 57 tokens to every session and 627 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
parcel-zoning-screen
Screen one Virginia parcel for its zoning district. Use when someone asks what a parcel is zoned, what a PIN or tax map number allows, whether a use is by-right, or what the zoning is at a Virginia address or point. It resolves the government first, then reads the parcel and zoning records, and it reports districts…
code-topic-walk
Reach a section of the Code of Virginia by topic when the reader has no citation, by walking the Code's own table of contents to it and then reading the section. Use when someone asks what state law says about zoning, annexation, FOIA, or any other subject without naming a section number. No public endpoint searches…
causallayer-mcp
This skill provides instructions for AI agents interacting with the CausalLayer Model Context Protocol (MCP) server. CausalLayer is a deterministic AI liability apportionment engine. It takes an incident report and mathematically proves which agent, vendor, or operator is liable based on counterfactual do-calculus…
korean-law
A legal research connection to South Korea's official law information service. It can retrieve statutes, court decisions, administrative rules, local ordinances, treaties, and legal interpretations, then check cited provisions against the source.
vendor-management
Manage ServiceNow vendors and contracts — corecompany vendor records, astcontract with renewal alerts, astcontractsla performance tracking, vendorriskassessment scoring, and vendor scorecards.
procure-quotazioni
Genera in serie procure alle liti (art. 83 c.p.c.) e lettere di quotazione compensi D.M. 55/2014 in DOCX per posizioni di recupero crediti, partendo da un Excel di posizioni o dai dati forniti, con rilevamento della fase processuale (monitorio, esecuzione forzata, opposizione a decreto ingiuntivo). Usa quando l'utente…