About sensor-calibration-workbench
A skill for Arduino projects where sensor readings are unreliable despite the sensor working. It guides calibration, evidence collection, coefficient persistence, and revalidation.
Key features
- 01Choose calibration type and workflow
- 02Collect reference and environmental evidence
- 03Review warm-up, drift, and saturation issues
- 04Plan coefficient storage and revalidation
Use cases
- 01Tune a load-cell factor
- 02Calibrate CO2, magnetometer, or color sensors
- 03Store calibration coefficients in EEPROM or flash
Install
Copy the command and run it in a terminal.
npx skills add https:/ / github.com/ wedsamuel1230/ embedded-engineering-skills --skill sensor-calibration-workbenchAfter installing, type/sensor-calibration-workbenchin Claude Code to confirm it loaded.
Bundled files
References and scripts installed along with SKILL.md.
references/calibration-flow.mdcommon-failure-patterns.mdpersistence-and-revalidation.mdDetails
From the project READMESensor Calibration Workbench
Use this skill when the sensor technically works, but the readings are not yet trustworthy enough for the project.
Resources
references/calibration-flow.md- end-to-end calibration workflow and evidence checklistreferences/common-failure-patterns.md- warm-up, scaling, drift, saturation, and environment mistakesreferences/persistence-and-revalidation.md- storing coefficients and deciding when recalibration is needed
When to Use
Use this skill when the request involves:
- volatile or implausible sensor readings
- "how do I calibrate this sensor?"
- load-cell factor tuning
- CO2, magnetometer, or color-sensor calibration
- storing calibration coefficients in EEPROM or flash
- deciding whether the problem is calibration, hardware, or environment
Do not use this skill when the sensor is not detected at all. That should route through hardware or bus bring-up first.
Workflow
- Confirm the measurement problem:
- unstable -> open
references/common-failure-patterns.md - offset or scaling error -> open
references/calibration-flow.md - values good once but bad later -> open
references/persistence-and-revalidation.md
- unstable -> open
- Identify the calibration class:
- one-point or zero-offset
- two-point scale calibration
- multi-orientation or environmental calibration
- Collect reference evidence before changing coefficients:
- known reference values
- warm-up state
- ambient conditions
- sample stability
- Decide how calibration values will persist and how revalidation will be triggered after reboot, firmware update, or field drift.
Core Rules
- Calibration without a known reference is guesswork.
- Warm-up and stabilization time are part of calibration, not a side note.
- Do not mix hardware-fault symptoms with coefficient-tuning symptoms.
- Store both the calibration values and enough metadata to know when they became stale.
Verification
- Confirm readings converge toward a known reference after calibration.
- Confirm the calibrated values stay stable across repeated samples.
- Confirm stored coefficients reload correctly after restart.
- If the project has operating thresholds, verify those thresholds against the calibrated output rather than the raw sensor value.
Integration
- Pair with
i2c-bringup-diagnosticianorcircuit-debuggerwhen the sensor is not yet electrically trustworthy. - Pair with
arduino-code-generatorwhen the user needs persistence or filtering code added to the sketch. - Pair with
field-power-and-connectivity-triagerwhen sensor behavior changes only off USB or under field power conditions.
Shared Output Contract
Use the shared Arduino skill contract: state assumptions, required tools and versions, implementation steps, tests/evidence by proof stage, known limitations, and recovery/security notes.