Create a branded LCV Ultrabrain reasoning session and optionally seed the first thought.

Canonical LCV reasoning gate for code work, branching, revisions, quality metrics, bias checks, confidence, and meta checkpoints. Protocol: state evidence and assumptions before conclusions; set step_type honestly (analysis, hypothesis, verification, conclusion...); attach evidence[] for any factual claim; branch when a materially different path exists and revise when an earlier step was wrong; adjust total_thoughts when scope changes instead of forcing a fit; only set next_thought_needed to false once a verification step has checked the conclusion. Close each step by asking: what am I missing or need to reconsider?

Update an existing thought with stronger evidence, confidence, risks, actions, or quality metrics.

Create or continue an Ultrabrain branch from a prior thought.

Merge one or more branch insights into a synthesis, best-evidence path, or decision note.

Check a reasoning session for evidence, alternatives, verification, bias, and closure gaps.

Analyze quality, confidence, bias counts, label counts, unresolved questions, and recommendations.

Render a reasoning session as a summary, linear chain, tree, markdown, or JSON.

Return session status, thought count, branches, and latest quality score.

Return recent normalized thought records for a session.

Export a session as markdown, text, or JSON.

Return aggregate session, thought, branch, quality, confidence, bias, mode, and step metrics.

List LCV Ultrabrain prompt templates for common engineering reasoning workflows.

Reset one Ultrabrain session, or all sessions when all_sessions is true.

Return runtime information for the local Ultrabrain MCP server, including version, release date, capabilities, tool surface, data directory, and active limits.