# Check Official Taiwan Weather Forecasts

Taiwan weather requests need reliable local forecasts with clear time periods. This skill guides CWA API queries for county and township forecasts using your authorization key.

## Install

```bash
npx skillstore add tahodev/cwa-weather
```

## Metadata

- Status: approved
- Slug: tahodev-cwa-weather
- Skillstore revision: r1
- Version status: missing
- Tree hash: a93ca99d2dffc221e82eb877a184c3cffdec405f1929e76a7d7c4d38e7220cea
- Author: tahodev
- GitHub username: tahodev
- License: MIT
- Repository: https://github.com/tahodev/baodao-skill/tree/3ce4b58e4525f5a66390e1236b6d38702790225a/cwa-weather
- Ref: 2aab20a3878daa05021884e089a7dd7739e08581
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, network, filesystem
- Quality score: 74
- Public page: https://skillstore.pages.dev/skills/tahodev-cwa-weather
- Manifest: https://skillstore.pages.dev/api/skills/tahodev-cwa-weather/manifest

## Capabilities

- Retrieve 36-hour county forecasts from the CWA F-C0032-001 dataset.
- Request multiple counties together or retrieve all counties.
- Interpret weather conditions, rain probability, minimum temperature, maximum temperature, and comfort indicators with forecast periods.
- Retrieve township forecasts from F-D0047 datasets and filter districts locally.
- Handle missing keys, authorization errors, timeouts, server errors, and invalid responses without inventing forecasts.
- Direct keyless forecast and warning requests to the separate taiwan-weather skill.

## Use Cases

- Plan a Local Trip: Review temperature ranges and rain probabilities for the next 36 hours before visiting a Taiwan county.
- Compare Work Locations: Compare forecast periods across multiple counties to support outdoor work scheduling.
- Prepare a District Weather Brief: Retrieve a township dataset, select the requested district, and summarize conditions with explicit forecast times.

## Prompt Templates

### Check One County

```
Check Taipei City weather for the next 36 hours using my configured CWA key. Include forecast periods, temperatures, and rain probabilities.
```

### Compare Two Counties

```
Compare Taipei City and New Taipei City forecasts for tomorrow. Match forecast periods and include rain probability, temperature ranges, and comfort indicators.
```

### Select a Township Forecast

```
Retrieve Taipei township forecasts and filter locally for Daan District. Summarize available periods and weather fields without exposing my configured authorization key.
```

### Handle Forecast Failures

```
Query county and township forecasts for Taipei using private temporary files. Validate responses, limit retries, and report failures without substituting historical data.
```

## Limitations

- Requires a valid CWA API key and network access; shell examples also require curl and jq.
- Township dataset IDs outside Taipei require catalog verification; township location filters may require local processing.
- Typhoon warnings, heavy-rain advisories, and earthquake reports are outside this skill; alternate feeds require the separate taiwan-weather skill.
- Historical test samples are not current forecasts, and documented API behavior may change.

## Best Practices

- Keep authorization keys outside replies, commits, process arguments, and logs.
- Use private temporary storage or direct parsing, and validate each API response before summarizing.
- Include forecast periods, verify dataset identifiers, and distinguish official forecasts from alternate sources.

## Anti Patterns

- Presenting historical examples or failed responses as current weather.
- Assuming township locationName parameters reliably filter every district.
- Using this forecast skill as a substitute for authoritative emergency warnings.

## Security Audit

- Audited at: 2026-10-13T02:46:28.454\+00:00
- Summary: Most alerts are false positives from Markdown backticks, official CWA links, dataset descriptions, and readable multilingual text. Predictable temporary files create overwrite and forecast-tampering risks, while command-line authorization URLs can expose API keys. No evidence found of prompt injection, malicious exfiltration, or concealed executable content.

## Stats

- Views: 0
- Downloads: 0
- Favorites: 0
- Popularity score: 0