Skip to content

SafeStreets MCP server

io.github.Poseidon-t/safestreetsPoseidon-tstreamable-httpv1.0.0Scanned

Walkability and pedestrian safety for any US address: scores, neighborhoods, deadly roads.

8 tools: 1 change things and 7 only read.

Tools: 0 delete, 0 irreversible, 1 write, 7 read.
  • Reads private data · 0
  • Reads text others write · 3
  • Can send data out · 0
  • 1 finding

Tools listed without sign-in. Remote posture

With Anlyon

1 tool here changes things. Govern it.

  • Anlyon lists the tools and adopts none.
  • Each tool you adopt gets approvals and receipts.
  • Your agent never holds the key.
Govern this server →

A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.

What the SafeStreets MCP server can do

The SafeStreets MCP server is published by the GitHub account Poseidon-t. It has 8 tools. It can create or change data (1 tool: score_many), but cannot delete data. 7 tools only read.

3 tools bring in text other people can write (get_walkability_score, compare_locations and score_many).

The fixed checks matched 1 finding (long-description). Each one comes with the text it matched.

Its remote endpoint lists its tools without sign-in, and its tools changed once since Anlyon Radar first saw them.

Last scanned Nov 8, 2026, version 1.0.0.

Tools

Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.

8 tools

Changes things 1Claude Code: ask
  • score_manyScore up to 10 US locations in one request.Reads text others writeNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Score up to 10 US locations in one request. Use this for a list: properties, branches, listings, sites. Each location counts as one call against your allowance, so a list of ten costs ten. A location that cannot be resolved is reported in errors rather than failing the request.

    Annotations the server declares

    • title: Score a list of locations
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
Only reads 7Show 7 read-only tools
  • compare_locationsScore 2 to 10 US locations and say how they differ.Reads text others writeNo findings

    Why it is in this group

    • The name "compare_locations" says "compare": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Score 2 to 10 US locations and say how they differ. Use this for "which of these is the most walkable", choosing between addresses, or any question that puts places side by side. Give addresses or coordinates, and a label for each if you want the results matched back to your own names. Returns every location ranked, the best and worst, the spread between them, and the gap on each component. The per-component gap is usually the answer: two places can reach the same composite for opposite reasons, one with errands close by and dangerous streets, the other the reverse. A location that cannot be resolved is reported in errors and the rest are still scored. Each location counts as one call against your allowance.

    Annotations the server declares

    • title: Compare locations for walkability
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • find_walkable_neighborhoodsNeighborhoods inside a US city, ranked most walkable first, precomputed.No findings

    Why it is in this group

    • The name "find_walkable_neighborhoods" says "find": a word for reading.
    Description, as the server wrote it
    Neighborhoods inside a US city, ranked most walkable first, precomputed. Use this for "which part of Chicago should I live in without a car". Returns up to 25 neighborhoods, each with its 0-10 score, tier and components. Covers 749 neighborhoods across the cities analysed so far, so an empty list means that city has not been broken down yet, not that it has no walkable areas.

    Annotations the server declares

    • title: Most walkable neighborhoods in a city
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
  • get_city_walkabilityWalkability score for a whole US city, precomputed.No findings

    Why it is in this group

    • The name "get_city_walkability" says "get": a word for reading.
    Description, as the server wrote it
    Walkability score for a whole US city, precomputed. Instant, no analysis. Use this for "how walkable is Houston" rather than scoring a point inside it. Returns the 0-100 composite, the same score as 0-10, the tier, and the four components. Covers 203 cities; a city not in the set returns null, which means not covered rather than not walkable.

    Annotations the server declares

    • title: City walkability score
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
  • get_dangerous_corridorsRoad corridors where pedestrians were actually killed, from federal crash records.No findings

    Why it is in this group

    • The name "get_dangerous_corridors" says "get": a word for reading.
    Description, as the server wrote it
    Road corridors where pedestrians were actually killed, from federal crash records. Source: NHTSA FARS 2022-2024 (PER_TYP=5, INJ_SEV in 4,6). Each corridor carries the road name, the city, the total pedestrian deaths and a per-year breakdown, so a claim can be checked rather than taken on trust. Give either a city, or a lat/lon with an optional radius. Use this when someone asks which roads are dangerous, where crashes happen, or wants evidence for a specific street. These are recorded deaths, not a model or a prediction: a road absent from the list had no recorded pedestrian fatality in the period, which is not the same as being safe.

    Annotations the server declares

    • title: Roads where pedestrians were killed
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
  • get_walkability_scoreScore a US location on how good and how safe it is to walk, from a street address or from latitude/longitude.Reads text others write1 finding

    Why it is in this group

    • The name "get_walkability_score" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Info

      The description is 2,533 characters long, over the 2,000 this check looks for.

      Matched text

      description ends: "...n driving directions, transit timetables, crime statistics, or property listings."

      What this check looks for
    Description, as the server wrote it
    Score a US location on how good and how safe it is to walk, from a street address or from latitude/longitude. Call this when someone asks whether a place is walkable, how safe it is to walk, whether they could live somewhere without a car, how a neighborhood compares to another on foot, or when evaluating a specific address or property for walkability or pedestrian risk. The server geocodes an address itself. Coordinates estimated from memory can land on a different place, which then gets a confident score of its own. When an address is given, the reply carries resolvedFrom with the address that was matched. Returns a JSON object: overallScore 0-100 composite displayScore the same score as 0-10, which is how SafeStreets presents it grade A to F tier plain label, e.g. "Very walkable", "Car-dependent" confidence 0-100, how much underlying data was available for this point reportUrl the full SafeStreets report for this point: map, metrics, PDF dataSources the sources that fed this specific result components four scored sub-dimensions, each with a 0-100 score and the individual metrics behind it: dailyReach errands reachable on foot streetSafety street design and pedestrian crash risk transitReach usable transit access walkingComfort shade, heat, air quality, surface plus networkDesign, which carries its metrics (network density, intersection density, block length, dead-end ratio) but reports score 0 at weight 0, so it does not contribute to the composite. Read its metrics, not its score. The keys safety, environmentalComfort and densityContext are backward-compatibility aliases of streetSafety, walkingComfort and dailyReach. They duplicate those objects rather than adding signal. The per-component metrics are the point: you can re-weight or rebuild the composite yourself instead of trusting one number. Coverage and limits: United States only. Pedestrian crash inputs are county-level, so two nearby points can share crash figures. Where the underlying crash record is missing, the safety component is reported as unavailable rather than guessed, and confidence drops accordingly. This does not return driving directions, transit timetables, crime statistics, or property listings.

    Annotations the server declares

    • title: Score an address for walkability
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • search_insightsSearch SafeStreets' published research on walkability, street design and pedestrian safety.No findings

    Why it is in this group

    • The name "search_insights" says "search": a word for reading.
    Description, as the server wrote it
    Search SafeStreets' published research on walkability, street design and pedestrian safety. Returns the article title, summary, URL and its published questions and answers, so a finding can be quoted and linked rather than paraphrased from memory. 32 articles, each built on named public sources. Use this when someone asks what the evidence says, wants a figure they can cite, or asks a general question about walkability that is not about one specific location.

    Annotations the server declares

    • title: Search SafeStreets research
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
  • search_neighborhoodsRanked neighborhood search across every US neighborhood (202,000+), for "where should I move" questions.No findings

    Why it is in this group

    • The name "search_neighborhoods" says "search": a word for reading.
    Description, as the server wrote it
    Ranked neighborhood search across every US neighborhood (202,000+), for "where should I move" questions. Use this when someone is relocating and has not picked a town yet: filter to a county set, a metro, a state, a home-price ceiling or a walkability floor, then rank by walkability, transit, zero-car share, demand or home value. Each row carries median home value, median income, price-to-income ratio, commute mode shares, zero-car household share, and where available year-over-year home value growth and property tax rate — so walkability can be read alongside what living there costs. Counties are addressed by 5-digit state+county FIPS codes because county names repeat across states (there are Delaware Counties in PA, OH, NY and IN). Example: Chester PA 42029, Delaware PA 42045, Montgomery PA 42091, New Castle DE 10003. walkabilityIndex is the EPA National Walkability Index on its 1-20 scale, with a national percentile beside it, and sort_by "walk" ranks on it. It is a different scale from the 0-100 SafeStreets score, which comes from scoring one address.

    Annotations the server declares

    • title: Search US neighborhoods
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false

This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.

Exfil surface

Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.

No private-data, untrusted-input, outbound path in this server alone: no tool is tagged private-data or outbound.

Reads private data0
Reads files, mail, messages, databases, secrets, calendars or repositories.
Reads text others write3
Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
Can send data out0
Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.

Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.

Remote posture

What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.

Tools listed without sign-in

Looked at Nov 20, 2026, 19:00 UTC, unchanged since Oct 22, 2026. No credentials were sent and no tool was called. What is looked at.

Reachable
yes
HTTP status of an anonymous initialize
200
tools/list served without credentials
yes
Tools that change things, listed without credentials
1
Protocol version
2025-06-18
WWW-Authenticate on 401
not observed
resource_metadata in WWW-Authenticate
not observed
RFC 9728 protected resource metadata
not observed
Authorization servers
not observed
Legacy HTTP+SSE transport
no
Dynamic client registration
not observed
Client ID metadata documents
not observed
  • WarnThe whole server

    The endpoint served tools/list without credentials, and the list has 1 tool that changes things (1 write). Whether a call also works without credentials was not tested: tools are never called.

    Matched text

    https://safestreets.streetsandcommons.com/mcp: tools/list returned 8 tools to an anonymous client, including score_many (write)

    What this check looks for

Findings

Matches of fixed checks, with the text they matched.

1 info
  • The description is 2,533 characters long, over the 2,000 this check looks for.

    Matched text

    description ends: "...n driving directions, transit timetables, crime statistics, or property listings."

    What this check looks for

History

Changes to its tool definitions and to its remote endpoint, newest first.

  1. Tools3 changed.~3 changedView diff

Install

Pinned to version 1.0.0. Values in angle brackets are yours to fill in.

claude mcp add --transport http safestreets https://safestreets.streetsandcommons.com/mcp
claude mcp add --transport http safestreets https://safestreets.streetsandcommons.com/mcp

Run in a terminal.

Claude Code permissions

Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.

.claude/settings.json
{
  "permissions": {
    "allow": [
      "mcp__safestreets__get_walkability_score",
      "mcp__safestreets__compare_locations",
      "mcp__safestreets__get_city_walkability",
      "mcp__safestreets__find_walkable_neighborhoods",
      "mcp__safestreets__search_neighborhoods",
      "mcp__safestreets__get_dangerous_corridors",
      "mcp__safestreets__search_insights"
    ],
    "ask": [
      "mcp__safestreets__score_many"
    ],
    "deny": []
  }
}

Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as safestreets. Review it before you use it.

Capability tags behind 3 rules
  • mcp__safestreets__get_walkability_score untrusted-input
  • mcp__safestreets__compare_locations untrusted-input
  • mcp__safestreets__score_many untrusted-input

Pin

Keep the decision: fail CI when the tools change, follow every change, or show a badge.

Pin in CI

Fail a build when the tools you reviewed are not the tools the server serves today.

curl -s "https://radar.anlyon.com/api/v1/servers/io.github.Poseidon-t/safestreets/verify?hash=be3aecc4485976da6a5e1ae7506f61d3a99315e176457ba41ef9930444857d92" | jq -e '.changed == false'

Pinned hash be3aecc44859, from Nov 8, 2026, 14:18 UTC.

Watch

Get told when its tools change: a new tool, a new description, a new class.

https://radar.anlyon.com/servers/io.github.Poseidon-t/safestreets/changes.xml

An Atom feed. Works in any feed reader or Slack's RSS app.

Badge

Anlyon Radar badge for io.github.Poseidon-t/safestreets
[![Anlyon Radar](https://radar.anlyon.com/badge/io.github.Poseidon-t/safestreets.svg)](https://radar.anlyon.com/servers/io.github.Poseidon-t/safestreets)

Source: MCP Registry · repository. Synced Oct 22, 2026, 11:17 UTC.

SafeStreets MCP server: questions and answers

Answered from the scan above, not from the publisher's description.

Is the SafeStreets MCP server safe?

No scan can promise that. What its definitions show: 8 tools, 1 of which can change, send or delete something, and 1 finding from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.

Can the SafeStreets MCP server delete data?

No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.

Can the SafeStreets MCP server do things that cannot be undone?

No tool is classed as irreversible: none of them sends, pays, publishes or merges, going by their definitions.

What tools does the SafeStreets MCP server have?

8 tools: 1 write, 7 read. score_many, get_walkability_score, compare_locations, get_city_walkability, find_walkable_neighborhoods, search_neighborhoods, get_dangerous_corridors and search_insights.

Does the SafeStreets MCP server need sign-in or OAuth?

Its remote endpoint listed its tools to an anonymous client, without sign-in.

Does the SafeStreets MCP server send data to outside services?

None of its tools sends, posts or publishes, or takes a recipient, channel or URL.

Has the SafeStreets MCP server changed its tools?

Yes, once since Anlyon Radar started tracking it, most recently on Nov 8, 2026. Every change is on this page and in its Atom feed.