A missing dependency does not always mean your HDL is wrong but an AI coding agent may still try to “fix” the source code. In this…

Sigasi

1,682 followers

A missing dependency does not always mean your HDL is wrong but an AI coding agent may still try to “fix” the source code. In this screencast, Copilot encounters dependency errors in a real HDL project. Instead of guessing, rewriting correct RTL, or burning tokens trying to infer the project structure, it asks Sigasi for deterministic, project-aware diagnostics through MCP. Sigasi identifies the actual problem: missing library definitions in the project.sigasi configuration. The agent updates the project setup - not the HDL - and Sigasi immediately analyzes the complete project again. This is what practical AI-assisted RTL development looks like: ✓ The coding agent does not treat every error as a code defect ✓ Sigasi understands the libraries, files, include directories, and dependencies ✓ The agent fixes the underlying project configuration automatically ✓ The HDL remains untouched when no source-code change is needed ✓ The corrected project becomes a reliable foundation for analysis and downstream tools Without deterministic project intelligence, an agent sees an unresolved reference and has to guess why it exists. With Sigasi, it can distinguish between broken HDL and a broken project setup... and apply the right fix. More code changes do not automatically mean more progress. Sometimes the safest change is no change to the RTL at all. The agent modifies. Sigasi grounds and validates. Engineers decide. ▶️ Watch our “Fixing Dependencies” in 2 minutes screencast: https://lnkd.in/eFMPJMz8 #FPGA #ASIC #RTL #SystemVerilog #VHDL #AgenticAI #EDA #HardwareDesign #MCP #SigasiInsights Fixing Dependencies with Sigasi MCP https://www.sigasi.com/screencasts/fixing_dependencies/