radaromega
THE weather source — all weather questions route here, from "is it gonna rain" to severe outbreaks.
Skill by Deibler
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/deibler-radaromega
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: sales, agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
- Runs as
- Python package
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/radaromega
What's inside
- Skill name
radaromega- Description
- THE weather source — all weather questions route here, from "is it gonna rain" to severe outbreaks. RadarOmega MCP for radar, satellite, model, MRMS and NDFD analysis, the convective-signs ladder that gives real lead time, alert watches, and meteorologist-grade annotated images and loop videos.
- License
- MIT
- Instructions
- 334 lines in SKILL.md · view on GitHub
Before installing
- Ships executable scripts
Files (9)
- SKILL.md
- signs.md · reference/
- cammap.py · scripts/
- cells.py · scripts/
- ground_truth.py · scripts/
- panel.py · scripts/
- radar_query.js · scripts/
- report.py · scripts/
- sounding.py · scripts/
SKILL.md (first lines)
# radaromega
RadarOmega is **the** weather workflow — not just severe weather. Current
conditions, "will it rain tomorrow", forecasts, radar, models, alerts, tropical,
winter, fire, marine, hydro, power outages: all of it starts here. Supplement
with `api.weather.gov` (via curl, with a User-Agent header) for text forecasts
and alert polling — never any other weather source.
Three companion documents, in the order you reach for them:
- **`reference/signs.md`** — the convective-signs ladder. Read it for any
"what is going to happen" question. Radar can only describe the present;
that file is how you get 30-90 minutes of honest lead time.
- **`skills/radaromega-pro`** — tool reliability, execute_js recipes, app
internals, data routing. Read WITH this skill before any real weather work.
- **`scripts/`** — `sounding.py`, `cells.py`, `ground_truth.py`, `panel.py`,
`report.py`, `cammap.py`, `radar_query.js`. Documented in radaromega-pro §5c/§6b.
## Reliability — what you can assume
- **Every tool self-heals.** No connect/discover step: each tool auto-attaches
and even auto-launches the app if it's closed (a cold first call takes ~10-45s
— send a quick "pulling up radar" first if the user is waiting). Tools never
hang; they error clearly.
- **Wedged engine auto-recovers.** If the app has been up a long time and the
model engine corrupts, `set_model` / `set_view_mode` detect it, restart the
app, and retry by themselves (~15-20s, result carries a `recovered:` note). If
`set_forecast_hour`/`set_model_product` say the app was restarted, just call
`set_model` again — NEVER drop into execute_js debugging for engine weirdness.
- "Stuck on a login or update screen" error → tell the user the app needs a
manual look; don't retry-loop.
- **Don't `sleep` in Bash.** Navigation tools wait for data themselves:
`change_radar_site` (default 5s settle), `change_radar_product` (3.5s), and
`capture_view(settle_ms: …)` covers the rest.
- **`get_detailed_state` lies about the mode.** It reports `viewMode: "RADAR"`
while satellite imagery is genuinely on screen. Trust `get_map_state`, the
product selector's option list, or your own eyes on a capture.
## Images & video — clean by default
- `capture_view` returns the **clean map** — data, warnings, your drawings, ZERO
app UI. That's what the user gets. Full-res goes to the sandbox for
`send_attachment`; by default only the saved path comes back (no inline image).
Pass `view: true` on captures you need to eyeball yourself — you get a
downscaled inline copy (don't re-capture to "get a better look" — the saved
file is sharp). Captures you're only delivering to the user don't need `view`.
- `capture_view(include_ui: true)` ONLY when you need to see the app's interface
to operate it. Never send include_ui shots to the user.
- `capture_loop(frames?, interval_ms?, format?)` — the animation as a clean mp4
(plays inline in iMessage) or gif. The single best way to show motion: an
approaching line, rotation, a model run, **a satellite loop of a building
cumulus field**. Works in RADAR, MRMS, MODEL and SATELLITE modes.
- `close_panels` — Escape + close-buttons when a settings panel covers the map.
## Don't spam the user
One reply = **the one or two visuals that tell the story**, with your read as the
caption. Run your full analysis pass silently (reflectivity, velocity, dual-pol,
satellite, whatever it takes — `capture_view(view: true)` on frames you are
reading yourself), then send the best frame plus maybe one loop. Eight separate
product screenshots is a slideshow, not an answer. Exception: the user explicitly
asks to see each product.
## Label like a meteorologist — always
Before sending an analysis image, annotate it (`clear_drawings` first, then draw,
then capture):
- Mark **what** and **where**: hail cores, rotation couplets, the user's
location, outflow boundaries, overshooting tops, threat polygons, motion vectors.
- Use real terminology and translate the so-what: "bow echo", "hook echo",
"training cells", "TVS", "gust front", "overshooting top", "above-anvil cirrus
plume" — then "that hook near Lititz is rotation, be on the basement side by 9:40".
- Captions state: what's happening, where it's going, when it arrives, what to do.
- `clear_drawings` and restore the home radar view when done with a remote event.
---
## Tool surface (verified arg names, product codes verified live 2026-09-03)
**Navigate** — `fly_to(lng, lat, zoom?)` (note `lng`), `fit_view(points, …)`
(use this instead of a guessed zoom before any capture), `measure_distance`.
**Modes** — `set_view_mode(RADAR|SATELLITE|MRMS|MODEL)`. The app's master
dropdown also carries `NDFD` and the model names directly, but the tool enum
does not; for NDFD see the recipe below.
### RADAR products — `change_radar_product(product)`
`HRF` reflectivity (also HRF2/3/4 for higher tilts of the *composite* product
set), `HVL` velocity (HVL2/3/4), `HSV` storm-relative velocity, `HSW` spectrum
width, `HCC` correlation coefficient, `DRF` differential reflectivity (ZDR),
`KDP` specific differential phase, `HCA` hydrometeor classification, `DAA` dual-pol
accumulation, `OST` one-hour storm total, `STP` storm total precip, `VIL`,
`ETP` echo tops, `WINTER_EXP`.
There is **no** `CC`, `ZDR`, `SHR`, `BRF`, `BVL` or `DVIL` — those are the codes
you would guess and the app rejects them. It validates against the live selector
and lists the real codes back at you; take the list it gives over this file.
**The elevation is fixed at the 0.5° base tilt.** There is no tilt control, so
BWER, storm-top divergence, descending reflectivity cores and ZDR *column depth*
are not available here. Vertical information comes from the derived products
(`VIL`, `ETP`) and from MRMS.
`change_radar_site(siteCode)` — waits for data itself. Pick the tower with the
**lowest beam** over the point in question, not the nearest by map distance.
### SATELLITE — works, and the old skill said it didn't
`set_view_mode("SATELLITE")`, then drive the product with **`change_radar_product`**
(yes, that tool — it writes the shared product selector, which in satellite mode
holds satellite codes):
- `GEOCOLOR-DATA` — day/night composite, the default look
- `NATCOLOR-DATA` — Day Land Cloud RGB. Ice cloud/snow renders **cyan**, water
cloud dull grey/white. Daytime only. The app's stand-in for Day Cloud Phase.
- `FIRETEMP-DATA` — fire temperature RGB
- `BAND01-DATA` … `BAND16-DATA` — all sixteen ABI bandsFrom Deibler/edmund-harness, MIT.
More from Deibler/edmund-harness
- 0.592score
iterative-design
Deibler
Use this skill whenever the user is iterating on a visual — flat graphics, logos, badges, t-shirt prints, hockey/sports graphics, or subtle photo edits across…
SkillAFreshness: A★ 1Claude Code +9 - 0.589score
hike
Deibler
Build a hike brief with a real topographic map, the traced route from OpenStreetMap, trailhead coordinates, and verified conditions.
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
deep-research
Deibler
Multi-agent deep-research orchestrator.
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
gifgrep
Deibler
Search Tenor/Giphy and send a GIF into the chat.
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
summarize
Deibler
TL;DR a URL, YouTube video, or local PDF/file via the `summarize` CLI (Gemini Flash).
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
cloudflare
Deibler
Headless browser automation via Cloudflare Browser Run — screenshot pages, render PDFs, extract markdown/HTML/links, scrape DOM elements with CSS selectors…
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
skycam
Deibler
The sky camera over the house (openskycam, camera sky01, Lancaster County PA), through the `mcp__openskycam__*` tools.
SkillAFreshness: A★ 1Claude Code +9 - 0.584score
agents
Deibler
Delegate long-running work (research, deep summaries, multi-step analysis) to detached sub-agents.
SkillAFreshness: A★ 1Claude Code +9 - 0.584score
fishing
Deibler
Ground any PA, MD, NJ, or DE fishing question in the local fishing dataset—waters, bass and other species, stocking, catches, boat access, motor rules, depth…
SkillAFreshness: A★ 1Claude Code +9 - 0.584score
teams
Deibler
Coordinate multiple sub-agents as one unit — parallel investigation, multi-stage pipelines, or side-by-side comparisons.
SkillAFreshness: A★ 1Claude Code +9 - 0.584score
trading-loop
Deibler
The autonomous trading decision loop for Quant.
SkillAFreshness: A★ 1Claude Code +9 - 0.584score
generation
Deibler
Comprehensive multimodal generation via OpenRouter — images (with reference photos), videos (text/image/reference-to-video), and spoken audio.
SkillAFreshness: A★ 1Claude Code +9
Signals
- Stars
- 1
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Sep 4, 2026 (1 month ago)
- Last release
- Oct 22, 2026 (5 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/Deibler/edmund-harness
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from Deibler
- 0.592score
iterative-design
Deibler
Use this skill whenever the user is iterating on a visual — flat graphics, logos, badges, t-shirt prints, hockey/sports graphics, or subtle photo edits across…
SkillAFreshness: A★ 1Claude Code +9 - 0.589score
hike
Deibler
Build a hike brief with a real topographic map, the traced route from OpenStreetMap, trailhead coordinates, and verified conditions.
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
deep-research
Deibler
Multi-agent deep-research orchestrator.
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
gifgrep
Deibler
Search Tenor/Giphy and send a GIF into the chat.
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
cloudflare
Deibler
Headless browser automation via Cloudflare Browser Run — screenshot pages, render PDFs, extract markdown/HTML/links, scrape DOM elements with CSS selectors…
SkillAFreshness: A★ 1Claude Code +9 - 0.587score
skycam
Deibler
The sky camera over the house (openskycam, camera sky01, Lancaster County PA), through the `mcp__openskycam__*` tools.
SkillAFreshness: A★ 1Claude Code +9
Cite this page
radaromega skill. Agent Mart. https://agentmart.dev/skills/deibler-radaromega (last updated 2026-10-24)