The 73 unscreened GND strokes are now flagged as boundary candidates, but the result stays overall_status UNKNOWN with engineering_sign_off false because EasyEDA Pro's beta PCB_PrimitivePoured.getAll does not expose whether current can actually reach. The author says plainly this is source geometry, not proof of copper contact.
Recurring agent failure issue in easyeda-mcp-pro (oaslananka)
What engineers report
A recurring agent failure problem in easyeda-mcp-pro (oaslananka) that mostly hits the electrical engineer.
Failure pattern: validation / geometry
Analyst note
From the Grok pain-radar digests, Nov 8-9 2026 (Nov 8 8:41 AM ET through Nov 9 6:41 AM ET digests plus the Haitch editions). Verified live 2026-11-09: issue/PR read; title and state confirmed.
What it costs a team
Modeled annual cost
$30,000
For a team of 5
Prototype delay
3 days
Modeled schedule slip per occurrence
Time lost
240 min
Per engineer per month, about 1 incidents
How the cost is modeled
240 min/month × 5 engineers × 12 months = 240 hours × $125/hour = $30,000
A modeled estimate with medium confidence. Swap in your own team size and loaded rate to size it for your team.
Engineering tools that address it
No engineering tool profile lists this error yet. Browse tools that connect to easyeda-mcp-pro (oaslananka).
Evidence
More recurring errors in easyeda-mcp-pro (oaslananka)
- A dual op-amp showed up as two schematic symbols, TL072CP.1 and TL072CP.2, and the read-only BOM bridge counted each one, reporting quantity 2 instead of 1 and risking an incorrect purchasing list. The fix counts a verified physical identity once and fails closed with BOM_IDENTITY_AMBIGUOUS when designators or IDs conflict.ECAD · impact 74
- A registered SCH_Event.addPrimitiveEventListener produced no callback during hand-drawn wire creation or wire dragging, and EasyEDA's own docs mark that API beta and not for production. Real schematics also hold repeated vertices and zero-length segments, and the old geometry helper treated every pin in the document as touching them, so a truly unwired part could not move. The fix compares two native readers over two samples 80 ms apart and refuses on any mismatch.ECAD · impact 74
- A user building a thin single-sided wearable says the standard sign-off is a look at EasyEDA's 3D view for body collisions, tall parts against the enclosure, and connector and flex exits, and right now an agent cannot help with this at all. The only proxy is a 2D capture of the component-shape layer, which misses parts without shapes and says nothing about height.MCAD · impact 74
- EasyEDA Pro 3.2.209 returns matching but stale wire lists from both getAll calls right after a wire is created or deleted, so a component move looks unwired and succeeds even though a connected pin would be disconnected or put on the wrong net. The fix stops trusting the batch list, tracks writes before they start, re-reads each wire by id, and refuses the move if it cannot confirm.ECAD · impact 74
- easyeda_schematic_batch_write and place_component report applied, but the part reads back Mirror false, AddIntoBom true, AddIntoPcb true. The bridge only applied rotation after create. Workaround until the fix in PR #805: place the part, then call modify with addIntoBom false. A part meant to stay off the BOM would have gone to the order.ECAD · impact 74
- easyeda-mcp-pro found that EasyEDA Pro returns an extensionless native filename, the project UUID, so the CPL and the Gerber zip landed on the same path in one artifact folder. The fix uses format-specific default names and tests both files exist with different contents.ECAD · impact 74
Sourced from public community reports. Costs and delays are modeled estimates, not measurements. Know a better source or a fix? Tell us at hello@productflo.io.