Generate Profile Skill

Overview

This skill defines the strict procedure for generating a BACnet Profile YAML.

Instructions

Step 1: Load Input Data

1. •Read the parsed result from /automation/temp/parsed-issue.json.
2. •Extract vendor, model, and device capabilities.

Step 2: Select Template

1. •Read the standard example: /examples/standard-profile/standard-temp-humidity-sensor.yaml.
2. •Smart Match: Based on keywords in bacnetMapping:
   • •"Binary", "Button", "Switch" -> /profiles/Moko/Moko-LW013SB.yaml
   • •"CO2", "Carbon" -> /profiles/Senso8/Senso8-LRS10701.yaml
   • •"Temperature", "Humidity" -> /profiles/Senso8/Senso8-LRS20100.yaml
   • •"Water", "Leak" -> /profiles/Senso8/Senso8-LRS20310.yaml
   • •Default -> /profiles/Senso8/Senso8-LRS20100.yaml

Step 3: Draft YAML Content

Compose the YAML strictly following the root-level structure:

• •codec: | Full JavaScript functions (Decode, Encode, decodeUplink, encodeDownlink).
• •datatype: Channel object mappings.
• •model, name, vendor, profileVersion (1.0.0), lorawan.

FORBIDDEN:

• •Do NOT use device: wrapper.
• •Do NOT use short strings for codec (e.g., "lorawan-1.0.3").
• •Do NOT use "datatype: json".

Step 4: Execution & Registration

1. •Construct the path: /profiles/{vendor}/{vendor}-{model}.yaml.
2. •Action: Call save_and_register(file_path, content). Do NOT use write_file.
3. •If file already exists, the tool will overwrite it.

Step 5: Verification

Call run_validation_script on the saved path.

• •If Success: Mark this skill as complete.
• •If Fail: Analyze the exact error message from the script output (e.g., "must have required property 'model'").
• •Fix the Draft YAML by ensuring it matches the root-level structure of the examples.
• •Repeat Step 4 and 5 (max 3 attempts).

Important Notes

• •All paths must be virtual (start with /).
• •save_and_register is MANDATORY for all output files.
• •The YAML must be valid and follow the examples exactly.