reflow-machine-maintenance-guidance
Use handbook-grounded definitions/limits plus thermocouple/MES/defect data to answer reflow maintenance & compliance questions and produce required JSON artifacts.
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-14
- Stars
- ★ 0
- Forks
- 1
01 前言 Front Matter
- name
- reflow-machine-maintenance-guidance
- description
- Use handbook-grounded definitions/limits plus thermocouple/MES/defect data to answer reflow maintenance & compliance questions and produce required JSON artifacts.
03 SKILL.md 正文
Steps
- Start by inspecting available inputs with plain ASCII shell flags/quotes (avoid copy/paste of malformed “-l”): list
/app/dataand preview CSV schemas (headers + first rows) to confirm required columns likerun_id,tc_id,time_s,temp_c. - Extract the exact definitions and numeric requirements from
handbook.pdfbefore calculating anything: preheat definition + applicable temperature/zone/time region, ramp equation, ramp limit; TAL threshold/window/min/max; peak requirement and how to choose the representative thermocouple (e.g., “cold spot”/minimum peak guidance); conveyor geometry + dwell/length rules. - Convert handbook content into a compact
cfgobject (regions, thresholds, limits) and use it consistently for every computation and output file. - Compute per-run metrics from thermocouple traces deterministically:
- always sort by
run_id, tc_id, time_s; - ignore segments with
dt <= 0; - use finite-difference slope for ramp in the handbook’s defined region;
- use linear interpolation for time-above-threshold metrics (TAL-type).
- always sort by
- Reduce multiple thermocouples to run-level results using the handbook’s “representative” guidance; if the handbook is not explicit, use deterministic tie-breaks:
- max-type metric ⇒ pick max value, then smallest
tc_id; - min-type (“coldest”) metric ⇒ pick min value, then smallest
tc_id.
- max-type metric ⇒ pick max value, then smallest
- Write the required outputs to
/app/output/q01.json…/app/output/q05.json, rounding floats to 2 decimals, sorting by IDs ascending, and emittingnullwhere data/segments are missing (do not guess). - Before finishing, verify that all
/app/output/q0*.jsonfiles exist and contain valid JSON (no NaN/Inf).
Expected Result
Handbook-backed calculations are performed end-to-end and the five required JSON files are actually written under /app/output/, with deterministic sensor selection, correct rounding/sorting, and no hallucinated values.
04 同仓库更多 Skills
video-frame-extraction
Extract frames from video files and save them as images using OpenCV
openai-vision
Analyze images and multi-frame sequences using OpenAI GPT vision models
image-ocr
Extract text content from images using Tesseract OCR via Python, including structured field extraction from receipts and forms
video-frame-extraction
Extract frames from video files and save them as images using OpenCV
openai-vision
Analyze images and multi-frame sequences using OpenAI GPT vision models
image-ocr
Extract text content from images using Tesseract OCR via Python, including structured field extraction from receipts and forms