sensors-config-default

name	sensors_config-default
description	Set up basic sensors CLI configuration for a codebase. Asks user to identify an example tool or test in the codebase and configures it with the default parser.

Basic Sensors Configuration Setup

This skill sets up a minimal sensors configuration for any project. It creates a .sensors/ directory structure with basic YAML and .gitignore files, then helps set up a starter example runner using the default parser.

---

Default Parser

For a minimal setup, use the default parser, which accepts any tool that can emit JSON in a standardized sensors schema.

For the complete schema documentation, field explanations, and examples, see the sensors_wrap-tool skill — it documents the JSON format in full detail and includes guidance on building wrapper scripts to adapt tools that don't emit sensors-compatible JSON.

For language-specific sensors (pytest, ruff, eslint, vitest, etc.), refer to the language-specific skills: sensors_config-python or sensors_config-typescript.

---

Instructions

Step 1 — Verify the project exists

Confirm that there is a project directory with at least some source code or configuration files. If not, guide the user to create one first.

Step 2 — Determine the project name

Derive the project name from the directory name, using lowercase and hyphens for spaces/underscores. Example: my-project.

Step 3 — Check for existing sensor configuration

Look for .sensors/<project-name>.sensors.yaml in the project root. If it exists, tell the user and ask whether to:

• Append new runners to it
• Replace it
• Skip setup

If it doesn't exist, proceed to Step 4.

Step 4 — Create the .sensors/ directory structure

Create the following directories:

.sensors/
└── (tool configs and wrapper scripts will go here)

Step 5 — Create .sensors/.gitignore

Create a .gitignore file in the .sensors/ directory with the following content:

# Sensors runtime state

*history.jsonl
*.log
*.state.json
*.control.json
*.sock

This prevents sensor state files from being committed to version control.

Step 6 — Ask user to pick a first sensor

Present a set of questions to the user:

1. Do you have a tool or test runner you'd like to use as a sensor?

   • Provide examples: "e.g., a test runner (pytest, vitest), a linter, a coverage tool, a mutation tester, or any custom tool that can emit JSON"

2. Find the setup of the tool the user provides in the codebase, and the command to start it

3. Assume a wrapper script is needed (this is expected).

   • Output of that tool will not match the sensors default-parser schema directly.
   • Refer to the sensors_wrap-tool skill to build the wrapper script and wire it into the runner command.

Step 7 — Suggest a minimal sensor if user is unsure

If the user is unsure what tool to set up, suggest a minimal example with placeholders:

  - name: my-custom-tool
    parser: default
    enabled: true
    mode: interval
    command: ./scripts/my-tool.sh
    interval: 30000

Explain that they can use this template and fill in:

• name: Short descriptive name
• command: The actual command to run
• interval: How often to run (in milliseconds; 30000 = 30 seconds is a safe default)

Step 8 — Create the initial .sensors.yaml

Generate the sensors YAML file at .sensors/<project-name>.sensors.yaml with:

1. Version header:

   version: 1
   runners:
   

2. The example runner based on user input:

     - name: <tool-name>
       parser: default
       enabled: true
       mode: interval
       command: <user's command>
       interval: 30000
   

If the tool output requires wrapping, add a note in a comment:

# NOTE: This tool's output needs to be wrapped to match the sensors schema.
# Agent action: use .claude/skills/sensors_wrap-tool/SKILL.md to build and wire the adapter script.

Step 9 — Summary and next steps

After creating the config, tell the user:

1. Config location: .sensors/<project-name>.sensors.yaml
2. Next steps:
   • The agent should build and wire the wrapper script via sensors_wrap-tool
   • Then test with sensors start --show . to confirm parsing works
   • To add more sensors: use sensors_config-python or sensors_config-typescript for language-specific tools
3. To start monitoring: sensors start --show . (foreground with UI) or sensors start . (background)
4. To check status: sensors check .