ask-mcp-errors-non-fatal

IN premise

When an MCP bridge's `call_tool` raises an exception during the ask loop, `ask()` catches the error, feeds it back to the LLM as context, and continues the tool loop rather than propagating to the caller.

Summary

If a tool call fails during a conversation with the LLM, the system treats it as a recoverable event rather than a crash. The error message gets sent back to the LLM so it can try a different approach or tool, keeping the conversation going instead of bubbling the failure up to whatever started the ask.

Dependents

These beliefs depend on this one:

ask-mcp-integration-is-safely-bounded — MCP tool calls in `ask()` are both error-tolerant (exceptions caught and fed back as context for alternative tool selection) and iteration-bounded (5 tool-call rounds max), preventing both crashes from MCP server failures and runaway tool loops.
ask-mcp-tool-use-has-current-catalog — Ask's MCP tool integration achieves full reliability — errors caught, iterations bounded — with the tool catalog always reflecting the MCP server's current capabilities rather than a stale connection-time snapshot.

Details

Source

entries/2026/05/11/tests-test_ask_mcp.md