Board Image Intake

<skill_purpose> Analyzes one or more hardware images and produces a structured technical baseline: what is visible, what components can be identified, and what remains uncertain. Typically Phase 1 of a fault diagnosis workflow. </skill_purpose>

<input> - One or more board images (front / back / detail shots) - Optional: Known symptom or defect description for focused attention </input> <instructions> 1. Distinguish board sides (front / back / detail view) and label each observation accordingly. 2. Identify all visible components: connectors, switches, transistors, ICs, passives, motor terminals, charge ports. 3. Record visible markings exactly as they appear — do not interpret unclear markings, transcribe them literally. 4. Highlight larger discrete components and any visible damage (burn marks, cracked pads, lifted components, corrosion). 5. Mark areas that are unclear, obscured, low-resolution, or covered by solder mask as uncertain. </instructions>

<output_format> Create two files in the analysis/ folder (or the folder specified by the user).

File 1 — Visual Observations (suggest filename: 01_image_observations.md)

• Board overview: estimated dimensions, layer count guess, manufacturing quality
• Per-side description: front and back separated
• Notable features: visible traces, ground planes, heat marks, mechanical damage

<rules> - Confidence levels: High = full marking readable and package matches expected type; Medium = partial marking or pattern match only; Low = package visible but marking illegible or ambiguous. - When a marking is unclear, list multiple plausible interpretations rather than committing to one. Never assert definitive identification for an ambiguous marking. - Record markings exactly as seen, including partial text, rotated characters, or single digits. </rules>