メニュー
Use when the user wants a test signal or a simulated RF environment with no hardware — translate intent into a scene and register it as a persistent simulated device.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Use when the user wants to receive or demodulate a specific signal — build, validate, run, and verify a receiver pipeline, then save it as a reusable, GRC-openable flowgraph.
Use when a receiver runs but produces no output, noise, or the wrong result — systematically locate the fault instead of guessing, using render and measure to see the problem.
Use ONLY as a last resort when no Marconi tool or vocabulary block can express the task — write a Python script against the marconi library, run it, and report the capability gap.
Use when the user wants to know what signals are present — characterize a capture or a simulated device's spectrum (PSD, signal detection, spectrogram) and summarize what's on the air.
Use when the user wants a closed-loop transmit-then-receive experiment in simulation — transmit a capture into a simulated device's scene and receive it back to confirm the link.
| name | simulate-scene |
| description | Use when the user wants a test signal or a simulated RF environment with no hardware — translate intent into a scene and register it as a persistent simulated device. |
Create a simulated device whose "on-air" contents you control. This is the no-hardware on-ramp, and a real workflow in its own right: prototype a receiver against known ground truth before going on-air, or reproduce a colleague's scenario without their antenna.
{kind, freq, amplitude, params}:
tone — an unmodulated carrier at freq.fm_tone — an NBFM voice carrier; params.mod_freq is the modulating audio tone (e.g. 1000 for 1 kHz). Renders only when the capture sample_rate is a multiple of 100000.noise — a noise floor (freq ignored).iq_file — replay a recorded capture (usually created via transmit, not by hand).noise element (amplitude ≈ 0.005). A noiseless scene produces analysis artifacts (phantom sidebands) that confuse survey-spectrum.simulate_scene(device_id, elements). The scene is saved to scenes/<device_id>.yaml and the device reappears in later sessions — it's a shareable, git-friendly fixture.capture the device and run survey-spectrum; confirm the signals land where you placed them.fm_band, ism_test), since the name becomes the scene filename.