A user of an Altium Designer agent bridge reports two bugs in one handler. Setting a parameter the project already has stops the bridge script behind an empty Altium "Error" box, and every session's calls fail until someone presses OK and restarts the loop. Setting a new parameter reads back "verified: true", but it is not in the .PrjPcb file on disk, because adding a parameter does not mark the project modified. Meanwhile other modified documents get saved that the caller never asked to save. The agent was told the edit worked. It did not.
Recurring agent failure issue in eda-agent (salitronic)
What engineers report
A recurring agent failure problem in eda-agent (salitronic) that mostly hits the electrical engineer.
Failure pattern: agent / false pass
Analyst note
From the Grok pain-radar digests, Nov 11-12 2026 (Nov 11 8:41 AM ET through Nov 12 5:41 AM ET digests plus the Haitch editions). Verified live 2026-11-12: 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 eda-agent (salitronic).
Evidence
More recurring errors in eda-agent (salitronic)
- "IPCB_LayerStack_V7.InsertInStackAbove / Below on a layer already in the stack leaves it where it is, with no error." The fix goes through the stack manager's own move. Because every move applies the manager's copy to the board, the tool now refuses to move while the two copies disagree, since a move would write the stale values over the board. A silent no-op API is the same failure as a false pass. The author had to build the comparison from #187 first just to make moving a layer safe.ECAD · impact 74
- "IPCB_LayerStack_V7.InsertInStackAbove / Below on a layer already in the stack leaves it where it is, with no error." The fix goes through the stack manager's own move. Because every move applies the manager's copy to the board, the tool now refuses to move while the two copies disagree, since a move would write the stale values over the board. A silent no-op API is the same failure as a false pass. The author had to build the comparison from #187 first just to make moving a layer safe.ECAD · impact 74
- Altium MCP bridge: running eda-agent relaunch from a terminal without the MCP server's EDA_AGENT_WORKSPACE overwrites a machine-global pointer file with the default folder, so the restarted Altium script serves that folder while the MCP server keeps sending requests to the custom one; then the diagnosis calls it a dead loop when the script is healthy, just somewhere else. The agent's context about where the design lives got silently swapped during recovery.ECAD · impact 74
- An Altium Designer 25.8 bridge for agents. The editor's Selection Filter also applies to a scripted `Prim.Selected := True`, and "for a type switched off there the assignment is silently refused." Every path that selects and then acts "does nothing for those types and still reports success," seen as "selected 51, removed 0" and select_objects returning a count with nothing selected. The fix turns the filter off around each programmatic selection and restores it.ECAD · impact 74
- An Altium Designer 25.8 bridge for agents. The editor's Selection Filter also applies to a scripted `Prim.Selected := True`, and "for a type switched off there the assignment is silently refused." Every path that selects and then acts "does nothing for those types and still reports success," seen as "selected 51, removed 0" and select_objects returning a count with nothing selected. The fix turns the filter off around each programmatic selection and restores it.ECAD · impact 74
- An Altium Designer 25.8 bridge for agents. The editor's Selection Filter also applies to a scripted `Prim.Selected := True`, and "for a type switched off there the assignment is silently refused." Every path that selects and then acts "does nothing for those types and still reports success," seen as "selected 51, removed 0" and select_objects returning a count with nothing selected. The fix turns the filter off around each programmatic selection and restores it.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.