功能描述

Austin CapMetro transit - real-time vehicle positions, next arrivals, service alerts, route info, and trip planning for buses and rail (MetroRail, MetroRapid...

使用说明 (SKILL.md)

CapMetro Austin Transit

Name: Capmetro Skill
Author: brianleach

Real-time Austin CapMetro transit data - vehicle positions, next arrivals, service alerts, and route information. No API key required.

When to Use

User asks about Austin bus or train schedules, arrival times, or delays
User asks "when is the next bus/train" or "is the 801 running"
User asks about CapMetro service alerts, detours, or disruptions
User wants to know where a bus/train currently is
User asks about nearby stops or route information
User mentions MetroRail (Red Line), MetroRapid (801/803), or any Austin bus route
User asks about CapMetro fares, how to ride, or general transit info

Data Sources

All feeds are open access, no API key required, hosted on the Texas Open Data Portal.

GTFS-RT (Real-Time) Feeds - Updated every 15 seconds

Feed	Format	URL
Vehicle Positions	JSON	`https://data.texas.gov/download/cuc7-ywmd/text%2Fplain`
Vehicle Positions	Protobuf	`https://data.texas.gov/download/eiei-9rpf/application%2Foctet-stream`
Trip Updates	Protobuf	`https://data.texas.gov/download/rmk2-acnw/application%2Foctet-stream`
Service Alerts	Protobuf	`https://data.texas.gov/download/nusn-7fcn/application%2Foctet-stream`

GTFS Static Feed - Route/Stop/Schedule data

Feed	Format	URL
GTFS Static (zip)	ZIP	`https://data.texas.gov/download/r4v4-vz24/application%2Fx-zip-compressed`

Implementation

Quick Start: Use the helper scripts

The scripts in this skill's scripts/ directory handle fetching, parsing, and presenting CapMetro data.

Script: `scripts/capmetro.mjs`

Main entry point. Supports these commands:

# Get current service alerts
node scripts/capmetro.mjs alerts

# Get real-time vehicle positions (optionally filter by route)
node scripts/capmetro.mjs vehicles [--route 801]

# Get next arrivals at a stop (by stop_id)
node scripts/capmetro.mjs arrivals --stop \x3Cstop_id>

# Get arrivals by searching stop name (uses best match)
node scripts/capmetro.mjs arrivals --stop-search "lakeline" --route 550

# Get arrivals filtered by direction/headsign
node scripts/capmetro.mjs arrivals --stop-search "downtown" --route 550 --headsign "lakeline"

# Get arrivals filtered by route at a stop
node scripts/capmetro.mjs arrivals --stop \x3Cstop_id> --route 801

# Search for stops by name or location
node scripts/capmetro.mjs stops --search "domain" 
node scripts/capmetro.mjs stops --near 30.4,-97.7

# List all routes
node scripts/capmetro.mjs routes

# Get route details including stops
node scripts/capmetro.mjs route-info --route 801

# Download/refresh GTFS static data (run periodically)
node scripts/capmetro.mjs refresh-gtfs

Setup: GTFS Static Data

On first use, run node scripts/capmetro.mjs refresh-gtfs to download and extract the static GTFS data (routes, stops, schedules) to ~/.capmetro/gtfs/. This only needs to be refreshed when CapMetro updates their schedule (typically quarterly or during service changes).

Key Route Reference

Route	Name	Type
550	MetroRail Red Line	Rail (Leander ↔ Downtown)
801	MetroRapid North/South	Rapid Bus (Tech Ridge ↔ Southpark Meadows)
803	MetroRapid Burnet/South Lamar	Rapid Bus (Domain ↔ Westgate)
1	N Lamar/S Congress	Local Bus
7	Duval/Dove Springs	Local Bus
10	S 1st/Red River	Local Bus
20	Manor Rd/Riverside	Local Bus
300	Oltorf/Riverside Crosstown	Crosstown Bus
325	Ohlen/Loyola	Crosstown Bus
985	Night Owl	Late Night Service

Tips for Users

Stop IDs can be found on CapMetro stop signs, in the Transit app, or by searching with the stops command
MetroRapid 801/803 have the most frequent service (every 10-12 minutes during peak)
MetroRail Red Line (550) runs Leander to Downtown Austin with limited frequency
Service alerts often contain detour information - check alerts before advising routes
Vehicle position data updates every ~15 seconds, so locations are near real-time

Error Handling

If a feed returns an error or empty data, inform the user that real-time data may be temporarily unavailable
The JSON vehicle positions feed is easier to parse and a good fallback if protobuf parsing fails
GTFS static data is required for stop names, route names, and schedule lookups - ensure it's been downloaded

Response Formatting

When presenting transit info to the user:

Lead with the most actionable info (next arrival time, active alerts)
Include route number AND name (e.g., "Route 801 MetroRapid")
Show times in 12-hour format with AM/PM
For delays, show both scheduled and estimated times
For vehicle positions, describe location relative to landmarks when possible
If there are active service alerts for the route the user asked about, always mention them

Fares Reference (as of 2025)

Fare Type	Price
Local / MetroRapid	$1.25
MetroRail	$3.50 (single)
Day Pass	$2.50
7-Day Pass	$11.25
31-Day Pass	$41.25

Payment via Umo app, tap-to-pay, or fare card. Free transfers within 2 hours.

External Endpoints

Endpoint	Data Sent	Data Received
`data.texas.gov/download/cuc7-ywmd/...`	None (GET only)	Vehicle positions (JSON)
`data.texas.gov/download/eiei-9rpf/...`	None (GET only)	Vehicle positions (Protobuf)
`data.texas.gov/download/rmk2-acnw/...`	None (GET only)	Trip updates (Protobuf)
`data.texas.gov/download/nusn-7fcn/...`	None (GET only)	Service alerts (Protobuf)
`data.texas.gov/download/r4v4-vz24/...`	None (GET only)	GTFS static data (ZIP)

All endpoints are open-access Texas Open Data Portal URLs. No API key, authentication, or user data is transmitted.

Security & Privacy

No credentials required - all data sources are open access, no API keys or tokens
No user data transmitted - requests are anonymous GET calls with no query parameters containing user info
Local storage only - GTFS static data is cached locally at ~/.capmetro/gtfs/; no data is written elsewhere
No telemetry - this skill does not phone home or collect usage data
Input handling - stop names and route IDs from user input are used only for local filtering, never interpolated into URLs or shell commands

Trust Statement

This skill only reads publicly available transit data from the Texas Open Data Portal. It does not access, store, or transmit any personal information. All network requests are read-only GET calls to open government data feeds.

安全使用建议

This skill appears to do what it claims: fetch public CapMetro GTFS/GTFS-RT data and present arrivals/vehicle positions. Before installing, confirm you trust the source (the GitHub repo listed) and that you're comfortable with these practical points: (1) it requires Node.js and the system 'unzip' binary and will run npm install (protobufjs) in the skill directory; (2) it will download and extract a GTFS ZIP into ~/.capmetro/gtfs and will overwrite files there when refreshing (the script uses unzip -o); (3) it only makes read-only GET requests to data.texas.gov (no credentials required). If you want extra caution, review the full scripts/capmetro.mjs file locally or run the skill in a sandboxed environment; optionally inspect the GTFS ZIP contents before extraction or modify the script to validate zip entries.

功能分析

Type: OpenClaw Skill Name: capmetro-skill Version: 1.0.3 The OpenClaw AgentSkills skill bundle for CapMetro transit appears benign. It adheres to its stated purpose of fetching and displaying public transit data from the Texas Open Data Portal. The `SKILL.md` and `scripts/capmetro.mjs` explicitly declare security and privacy practices, which are consistent with the code's behavior: no credentials required, no user data transmitted, and data is cached locally in `~/.capmetro/gtfs/`. The use of `execFileSync` for `unzip` to extract GTFS static data is a standard practice, with arguments passed as an array, mitigating shell injection risks. User inputs are used for internal filtering and searching, not for constructing shell commands or external URLs. There is no evidence of data exfiltration, persistence mechanisms, or malicious prompt injection attempts against the AI agent.

能力评估

✓ Purpose & Capability

Name/description (CapMetro transit data) match the included files and runtime behavior. The script fetches GTFS and GTFS-RT feeds from the Texas Open Data Portal, parses protobuf/JSON feeds, and uses GTFS static files for stop/route lookups. Required binaries (node, unzip) are reasonable for a Node.js script that downloads and extracts a ZIP.

ℹ Instruction Scope

SKILL.md and README instruct the agent/user to run npm install and to run the included node script and refresh-gtfs. The runtime instructions are narrowly scoped to fetching/parsing CapMetro feeds and caching GTFS in ~/.capmetro/gtfs. The skill documentation and the script explicitly state there are no env vars or other system-wide reads. Note: the skill will write files under the user's home directory (~/.capmetro/gtfs) and will make outbound GET requests to data.texas.gov; this is expected but worth noting.

ℹ Install Mechanism

There is no automated registry install spec, but SKILL.md (and the README) instructs running npm install (or an inline npm install --prefix $SKILL_DIR protobufjs). The package.json declares protobufjs as a dependency and package-lock.json references public npm packages — this is proportionate. The script uses the system 'unzip' via execFileSync to extract the downloaded GTFS ZIP; the feed URL is data.texas.gov (a well-known public host). Minor caution: extracting remote archives with system unzip (-o) can overwrite files; the code writes into ~/.capmetro/gtfs and does not validate archive contents for path traversal or unexpected filesystem writes, which is a generic risk when extracting remote archives but consistent with the skill's purpose.

✓ Credentials

The skill declares no required env vars or credentials and the code does not read credentials. All external endpoints are the Texas Open Data Portal URLs listed in SKILL.md and the script comments. There are no unexpected secrets requested.

✓ Persistence & Privilege

The skill is not always-included and can be invoked by the user or the agent. It persists cached GTFS files under the user's home directory (~/.capmetro/gtfs) which is reasonable for this use. It does not modify other skills or global agent configuration.

版本历史

v1.0.3

Fix vehicles command showing zero results (camelCase routeId field)

v1.0.2

List files explicitly in manifest to satisfy ClawHub scanner

v1.0.1

declare unzip as required binary dependency

v1.0.0

Initial release

元数据

Slug capmetro-skill

版本 1.0.3

许可证 —

累计安装 0

当前安装数 0

历史版本数 4

常见问题

Capmetro Skill 是什么？

Austin CapMetro transit - real-time vehicle positions, next arrivals, service alerts, route info, and trip planning for buses and rail (MetroRail, MetroRapid... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 728 次。

如何安装 Capmetro Skill？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install capmetro-skill」即可一键安装，无需额外配置。

Capmetro Skill 是免费的吗？

是的，Capmetro Skill 完全免费（开源免费），可自由下载、安装和使用。

Capmetro Skill 支持哪些平台？

Capmetro Skill 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。

谁开发了 Capmetro Skill？

由 Brian Leach（@brianleach）开发并维护，当前版本 v1.0.3。

Capmetro Skill