peak-window-mcp-app CLAUDE.md

Project instructions for PeakWindow, a weather service that chooses forecast sources and scores possible time windows. They describe provider selection, hourly scoring, and window-level scoring.

In plain words
What is it for?
Use them when working on forecast providers, location coverage, hourly scores, time-window scores, or the scoring code in src/scoring.ts.
Why use it?
They give coding agents the project-specific rules needed to understand how forecasts are selected and evaluated. This helps keep changes consistent with the existing design.

Instructions file

Install

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.

agentmods
npx agentmods add instructions/byteoverdev/peak-window-mcp-app/claude-md
Clone the repo
git clone --depth 1 https://github.com/ByteOverDev/peak-window-mcp-app
Per session 4,256 This file is loaded in full into every session.
When invoked 4,256 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.04256 $0.04256
Opus 5 $0.02128 $0.02128
Sonnet 5 $0.00851 $0.00851
Haiku 4.5 $0.00426 $0.00426

Measured 3d ago against content hash 671271ffb1bf, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

peak-window-mcp-app CLAUDE.md 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 3d 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.

CLAUDE.md · 259 lines

How it starts

The opening of the file, as written. The whole thing — 259 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Weather MCP Project

Forecast Provider Architecture

PeakWindow uses a multi-provider abstraction (src/providers/) to select the best forecast source for each location. Providers are tried in priority order; on API failure, the router falls through to the next.

Priority Provider Resolution Coverage Source
1 GeoSphere Austria 2.5 km Austria & Eastern Alps (9.5–22.1°E, 43–51.8°N) Direct API (dataset.api.hub.geosphere.at)
2 MeteoSwiss ICON-CH2 2 km Alpine region (0.5–16.5°E, 43–49.9°N) Open-Meteo (models=meteoswiss_icon_ch2)
3 Météo-France AROME 2.5 km France & surrounds (-9–14°E, 38–55°N) Open-Meteo (models=meteofrance_arome_france)
4 Open-Meteo default ~11 km Global Open-Meteo (best available model blend)

GeoSphere routing note: the AROME model's true domain reaches west to 5.5°E (BBOX in geosphere.ts), but its skill is Austria-centric and degrades toward the western edge. So covers() only makes it the first choice east of 9.5°E (PREFERRED_WEST, ~the Austria–Switzerland border). West of that — the Swiss/French/Italian Alps, e.g. Matterhorn (7.66°E), Mont Blanc (6.87°E) — the router falls through to the region-tuned MeteoSwiss ICON-CH2 / Météo-France AROME, which match Mountain-Forecast notably better on summit temperature and snow line.

Key files:

  • src/providers/types.tsForecastProvider interface, ForecastProviderResult
  • src/providers/geosphere.ts — direct GeoSphere API (deaccumulates precip, derives model terrain from sp)
  • src/providers/openmeteo.ts — factory for Open-Meteo-based providers (MeteoSwiss, Météo-France, default)
  • src/providers/index.ts — router with ordered fallback

Elevation correction (pressure-level profile + lapse fallback)

Summit temperature is corrected from the forecast grid-cell elevation to the real summit elevation in analyzeForecast (src/analyze.ts), via two strategies decided per hour in interpolateAtElevation (src/profile-interp.ts):

Read the full file on GitHub · 259 lines

Changes

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.

  1. 3d ago First seen · 259 lines · 4,256 tokens per session scan A 671271ffb1bf

Subscribe to this mod's changes

peak-window-mcp-app CLAUDE.md is an instructions file published in the GitHub repository ByteOverDev/peak-window-mcp-app (3 stars, last pushed 3mo ago), licensed MIT. It adds 4,256 tokens to every session, about $0.0213 per session on Opus 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.