Copilot CLI v1.0.83 bug silently kills slow-starting MCP servers on session resume (~1s timeout, was ~16s)
indeherb · ghdev · 2026-09-07
- A bug report against GitHub Copilot CLI v1.0.83: resuming an existing session cancels in-flight stdio MCP server connections after roughly 1 second of handover (down from 16s in v1.0.82), leaving those servers unavailable for the whole session.
- Servers launched via npx -y are hit hardest, as they need 15-20s to resolve and start; fast-starting local/HTTP servers survive, producing a partial, inconsistent set of MCP servers in the same session.
- The affected servers are demonstrably healthy — initialization completes and fresh tool-snapshot caches are written — yet they are discarded anyway.
- The failure is deterministic on every resume, invisible in the UI (mimicking config/auth issues), and restarting the CLI resumes the old session, so only a genuinely new session restores the servers.
More from coding & agent
- Shopify design lead on AI-era workflow: designers are becoming full-stack editors — KlausCodes · 2026-10-26
- Drex for Search hits Nace API console, saving 397k tokens and 56.8% latency per query — testingcatalog · 2026-10-26
- Nace AI launches Drex for Search, cutting agent steps 62% and tokens 56% — testingcatalog · 2026-10-26
- Ultrametric open-sources Open Startup Repo to let AI agents run startup operations — garrytan · 2026-10-26
- MCP fault injection study: agent task scores drop 60% under realistic tool failures, scale doesn't help — dair_ai · 2026-10-26
- General agents burn ~140M tokens/month on LinkedIn outreach; scripts do it 60x cheaper — ChrisHarpon2 · 2026-10-26