← Back to Skills Marketplace
S2-SP-OS Wave/Spectrum Radar
by
MilesXiang
· GitHub ↗
· v2.0.3
· MIT-0
127
Downloads
0
Stars
0
Active Installs
4
Versions
Install in OpenClaw
/install s2-spectrum-perception
Description
S2-SP-OS Spectrum Radar. Pure, passive spatial perception using GPIO/UART with explicit OS environment consent (S2_PRIVACY_CONSENT) and quantized biometrics....
Usage Guidance
This package mostly implements a local, passive radar client and enforces an explicit S2_PRIVACY_CONSENT env var, which is reasonable. However, two issues merit caution before installing: (1) the edge setup doc mentions fast GPIO triggers (implying actuation possibility) while the code and SKILL.md repeatedly assert 'passive only' — ask the author to clarify whether the hardware wiring or GPIO pins can be used to trigger other devices; (2) AGENT-EXAMPLES.md encourages the Agent to autonomously call actuators/other sensors based on radar output. If you plan to allow autonomous skill invocation, restrict or audit cross-skill policies so the Agent cannot trigger actuators without separate explicit approval. Practical steps: run the Python script in an isolated environment first (no /dev/tty or GPIO attached) to inspect outputs; review/grep the code for any network calls (none present in provided file); avoid setting S2_PRIVACY_CONSENT=1 globally until you are comfortable (set it only in a controlled session); and ask the maintainer to remove or rewrite examples that instruct the Agent to take autonomous actions or to explicitly document how GPIO trigger wiring is intended to be used. If the maintainer confirms that GPIO pins are read-only and no actuator interfaces exist, the coherence concern is resolved and risk decreases.
Capability Analysis
Type: OpenClaw Skill
Name: s2-spectrum-perception
Version: 2.0.3
The skill is a passive spatial perception tool designed to interface with mmWave radar hardware via GPIO or UART. It includes a safety gate requiring the environment variable 'S2_PRIVACY_CONSENT=1' and implements a 'privacy quantization' logic in spectrum.py to convert raw biometric data into semantic states (e.g., 'Normal') rather than exposing raw values. There is no evidence of data exfiltration, malicious execution, or harmful prompt injection; the instructions and code are consistent with its stated purpose of providing sensory data for home automation.
Capability Assessment
Purpose & Capability
Name/description, required binary (python3), and dependency (pyserial) align with a GPIO/UART radar client. However, the edge-setup guide references 'GPIO Fast-Trigger Wiring' for 'instant zero-latency smart home triggers', which contradicts repeated claims in SKILL.md and spectrum.py that the sensor is 'passive' and performs 'NO cross-device triggers.' This inconsistency is unexplained.
Instruction Scope
SKILL.md instructs running spectrum.py and explicitly requires S2_PRIVACY_CONSENT (consistent). But AGENT-EXAMPLES.md contains sample agent reasoning that tells the Agent to autonomously call other perception/actuation skills (e.g., s2-light-perception, s2-acoustic-perception). SKILL.md also asks the Agent to 'decide what to do next based on your system prompts and examples' — this gives broad discretionary authority to act on sensor outputs and promotes cross-skill actuation despite claims of sensor-only scope.
Install Mechanism
No install spec downloads arbitrary code; this is instruction-only plus an included Python file. The only declared runtime dependency is pyserial (normal for serial/UART hardware). No network downloads or extract steps are present.
Credentials
Only one required environment variable (S2_PRIVACY_CONSENT) is declared and enforced by the script. No unrelated credentials, config paths, or secrets are requested.
Persistence & Privilege
always:false and no install-time persistence are appropriate. However, because the skill provides examples that encourage the Agent to invoke other skills or actuators autonomously, the combination of autonomous invocation (default enabled) plus these examples increases operational risk if the Agent is allowed to act without policy controls.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install s2-spectrum-perception - After installation, invoke the skill by name or use
/s2-spectrum-perception - Provide required inputs per the skill's parameter spec and get structured output
Version History
v2.0.3
- Now explicitly declares the required OS environment variable (S2_PRIVACY_CONSENT) in metadata for sandbox environments.
- Clarifies pure, passive sensing only: the skill no longer assumes or executes any cross-device actions or triggers.
- Updates documentation to stress that exact biometric data is destroyed at the hardware edge and only quantized data leaves the sensor.
- Refines agent instructions: the skill outputs perceptual data only; all decisions and actions are now delegated to the agent.
v2.0.2
- Now requires explicit OS-level consent via the S2_PRIVACY_CONSENT=1 environment variable (replaces CLI magic string tokens).
- Removed references to "consent-token" from command examples; usage now simpler and centralized.
- Maintains strict semantic-only health data output—no exact BPM logged or output.
- Clarified privacy and operational instructions for AI agents.
v2.0.1
s2-spectrum-perception 2.0.1 – Major Privacy & Interface Redesign
- Added strict health data quantization: system outputs only semantic ranges (never exact BPM) to maximize biometric privacy.
- Replaced raw FMCW/FFT compute with secure Advanced UART protocol for vital sign parsing; commercial hardware MCU is always used.
- Now supports both GPIO (for presence/motion) and UART (for vital sign energies) hardware modes.
- Updated dependencies: replaced numpy/scipy with pyserial for serial communication.
- Enhanced instructions detail explicit consent handling and privacy by design.
v1.0.0
s2-spectrum-perception 1.0.0
- Initial release of S2-SP-OS Wave/Spectrum Radar skill.
- Supports two modes: direct hardware pin reading for instant presence detection, and local raw FMCW signal processing for vital signs (breath and heart rate).
- Uses numpy and scipy for local Edge-Compute (FFT, phase unwrapping, filtering).
- Maps sensor data to 4-sqm spatial voxels for precise localization.
- Provides clear agent instructions for occupancy, vital-sign scan, and multimodal escalation.
Metadata
Frequently Asked Questions
What is S2-SP-OS Wave/Spectrum Radar?
S2-SP-OS Spectrum Radar. Pure, passive spatial perception using GPIO/UART with explicit OS environment consent (S2_PRIVACY_CONSENT) and quantized biometrics.... It is an AI Agent Skill for Claude Code / OpenClaw, with 127 downloads so far.
How do I install S2-SP-OS Wave/Spectrum Radar?
Run "/install s2-spectrum-perception" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is S2-SP-OS Wave/Spectrum Radar free?
Yes, S2-SP-OS Wave/Spectrum Radar is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does S2-SP-OS Wave/Spectrum Radar support?
S2-SP-OS Wave/Spectrum Radar is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created S2-SP-OS Wave/Spectrum Radar?
It is built and maintained by MilesXiang (@spacesq); the current version is v2.0.3.
More Skills