From marconi
Characterizes RF spectrum captures using PSD, signal detection, and spectrogram analysis to identify and summarize signals on the air.
How this skill is triggered — by the user, by Claude, or both
Slash command
/marconi:survey-spectrumThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Characterize what's present in a capture, then summarize it for the user. This is the first step of almost every RF task ("what's here?").
Characterize what's present in a capture, then summarize it for the user. This is the first step of almost every RF task ("what's here?").
capture(device_id, center_freq, sample_rate, duration). From a file the user brings: load_capture(path). Use the returned path for everything below.psd(capture_path) → noise floor + strongest peaks. find_signals(capture_path) → candidate signals (center_freq, bandwidth, snr_db).spectrogram(capture_path) and read the image. find_signals is a first pass over a 1-D PSD; the spectrogram shows time/frequency structure (bursts, drift, multiple carriers) that the list can miss. If you see on/off activity, run detect_bursts(capture_path) for structured burst timing (start, duration, power).measure. A wide signal (e.g. an FM carrier) with a low noise floor can fragment into several find_signals detections. For any wide or clustered group, call measure(capture_path, center_freq=...) to get the true 99% occupied bandwidth, and trust the spectrogram over the raw detection count.noise element (amplitude ≈ 0.005).measure result as a real signal only above ~8 dB SNR; below that you're likely measuring noise.measure when you need a real bandwidth number.npx claudepluginhub yoelbassin/gr-mcp --plugin marconiSystematically debugs DSP faults (silent audio, noise, wrong pitch) in GNU Radio receivers by inspecting spectrograms and checking one parameter per iteration.
Analyzes digital/analog captures from Saleae Logic MSO devices, decoding protocols like UART, SPI, I2C from exported binary files. Use for CTF challenges, hardware reverse engineering, or protocol analysis.
Loads, generates, lists, inspects, caches, and visualizes vibration signals (CSV, TXT, NPY, WAV, MAT, Parquet) in predictive-maintenance-mcp server's in-memory repository.