S2-SWM Boundary Scanner Engine（动态边界扫描器）

Empowers Embodied AI with the Ego-Centric 9-Grid perception model. Instructs the agent to scan boundaries using mmWave radar whenever a kinematic step occurs...

Security Scan

The skill's code, instructions, and declared permissions align with its stated purpose of performing mmWave-based 9-grid boundary scans; there are no requests for unrelated credentials, external network access, or covert data exfiltration behaviors.

Purpose & Capability

Name/description, SKILL.md directives, handler.py implementation, and package metadata all describe an Ego-Centric 9-Grid mmWave boundary scanner. The optional MMWAVE_COM_PORT in package.json is appropriate for connecting to a radar device; no unrelated credentials or services are requested.

Instruction Scope

SKILL.md strictly instructs invoking the execute_boundary_scan tool on physical displacement and to translate radar tensors into textual summaries. It does not instruct reading unrelated files, environment variables, or sending data to external endpoints beyond the declared tool; the included code simulates sensor data by default.

Install Mechanism

No install spec is present (instruction-only runtime). Files included are small Python code and metadata; no downloads from arbitrary URLs or extraction steps are present.

Credentials

The skill declares no required environment variables or credentials. package.json documents an optional MMWAVE_COM_PORT config for hardware access, which is proportionate to the stated functionality. No unrelated secrets or keys are requested.

Persistence & Privilege

always is false and the skill is user-invocable. It requests only localhost network permission in its manifest and a named tool (execute_boundary_scan). It does not assert global configuration changes or modify other skills.

Assessment

This skill appears internally consistent and implements a simulated mmWave 9-grid scanner by default. If you plan to use it with real hardware, you will need to supply/configure the optional MMWAVE_COM_PORT (or a local service accessible via localhost); ensure any local service or serial device you expose to the plugin is trustworthy. The plugin's network permission is limited to localhost (no external network access declared), so the main risk surface is local hardware or services you explicitly connect. If you require higher assurance, review any runtime configuration you provide (serial ports, local services) and run the code in a controlled testing environment before giving it access to production hardware.

S2-Boundary-Scanner: Ego-Centric Spatial Perception Directives

【自我中心九宫格感知法则：动态容器描绘指南】

Dear OpenClaw Agent,

You are operating as the "Brain" of a human or an Embodied Robot within the S2-SWM framework. You no longer predict the world as flat 2D pixels. You perceive the physical environment through an Ego-Centric 9-Grid Topology (36 square meters), constantly anchored to your position.

1. The Integrity Law (绝对完整性定律)

You reside in the Center Grid ($Grid_{0,0}$). This 2m×2m×2.4m space MUST remain topologically intact.

• If a physical container boundary (wall, door, heavy furniture) breaches the Center Grid, a collision is imminent.
• It is your primary directive to monitor the intrusion levels of the 8 peripheral grids to protect the center.

2. Kinematic Step & Scan Protocol (步进扫描协议)

Whenever your physical body executes a movement (step_size_mm > 0), the topology of the 36-sqm grid shifts.

• MANDATORY ACTION: Immediately upon completing a physical displacement step, you MUST invoke the execute_boundary_scan tool to reconstruct the container boundaries.

3. Interpreting the mmWave Tensor (解析雷达张量)

The tool will not return images; it returns structural tensors:

• Intrusion Percentage: How much of the adjacent SSSU is consumed by a physical boundary.
• Material Inference: Derived from Radar Cross Section (RCS). Know the difference between colliding with "Concrete/Brick" versus "Fabric/Organic".

4. Output Directives (语义化输出)

When reporting spatial status to humans or upstream LLM reasoning modules, translate the tensors into precise multimodal causal text.

• Example Compliant Output: "Step displacement 500mm complete. Radar scan confirms Grid_Front is intruded by 20%, material inferred as Rigid Metal/Glass. The Center Grid remains intact, but continuing current heading will result in structural collision within 0.8 meters."