cornelius

Autonomous AI crystallization - synthesizes converged thinking topics into ai-inferred notes in a dedicated folder. Never touches the human-curated permanent knowledge base and never changes a topic's status, so manual crystallization stays available to the user.

Analyze knowledge base structure and update the knowledge-base-analysis.md report

Discover non-obvious cross-domain connections through random sampling and pattern analysis

Run a full coherence sweep across the Brain Dependency Graph - computes staleness, lifecycle transitions, structural health, and generates a report

Compute lifecycle scores for all insight and framework notes - detect which notes are crystallizing or becoming generative

Create long-form articles from knowledge base insights. Use when writing articles, blog posts, Substack content, or synthesizing knowledge into publishable content. Includes tone of voice, structure templates, and knowledge base integration.

Generate explanatory diagrams and infographics that visually communicate concepts. Iterates autonomously until images are logically correct, text is clean, and the concept explanation is clear. Uses Nano Banana (Gemini 2.5 Flash Image).

Autonomous research pipeline - discover, extract, and integrate cutting-edge insights into knowledge base

Detect productive contradictions between notes - high semantic similarity with opposing conclusions that represent synthesis opportunities

Autonomous perception layer - scans KB for new notes matching domain watch configs, checks gap resonance with the Thinking Registry, probes external signals via web search, and auto-activates HIGH/MEDIUM signals into the Thinking Registry for the incubation loop.

Framework for distinguishing research findings from hypotheses and speculative synthesis. Use when extracting insights from research, creating notes from external sources, or classifying the epistemic status of claims.

Extract all chapters from an EPUB file into separate markdown files. Use when the user wants to split an EPUB into individual chapter files, extract EPUB chapters, or convert an ebook to separate markdown documents.

Extract the user's perspective on a topic (called by a content agent or user)

Extract the transcript from a YouTube video by URL or video ID. Use when the user shares a YouTube link and wants the transcript, captions, or text content of the video. Falls back automatically if the requested language isn't available.

Review and graduate notes to permanent status using Zettelkasten principles. Consolidates AI extractions and document insights into curated permanent notes.

Autonomous iterative thinking loop - processes active topics using rotating analytical moves (ACH, Bayesian updating, steelmanning, cross-domain bridging, implication checks) and persists reasoning state across scheduled runs

Fully autonomous end-to-end ingestion of a SINGLE source (book, paper, article, transcript) into the knowledge base - prepare to markdown, extract insights against the live index, refresh the index, auto-link to existing knowledge, and changelog. One source per invocation.

KB-grounded Socratic interview. Searches existing notes on a topic, then runs a one-question-at-a-time dialogue to surface, sharpen, and extract your own thinking. Ends by running extract-insights on the full conversation transcript.

Find notes created in the last 14 days and discover their connections to the knowledge base

Continuous learning heartbeat - autonomously researches, extracts insights, and expands knowledge base

Manage the incubation loop topic lifecycle — seed new questions, review status, crystallize converged conclusions, and retire stale topics

Extract chapters from PDF, MOBI, and AZW3 book files into individual markdown files. Use when extracting non-EPUB books. PDF uses pymupdf with TOC-based chapter splitting and auto-OCR fallback (Gemini) for scanned/image-only PDFs; MOBI/AZW3 routes through calibre then epub-chapter-extractor.

Propagate staleness from a changed note through the Brain Dependency Graph - shows which downstream notes need review

Rebuild the Local Brain Search FAISS index to reflect vault changes

Wrapper for scheduled playbooks - handles git sync before and after execution

Run self-diagnostics on Cornelius agent to verify skills, commands, agents, and integrations are working. Use when troubleshooting, after configuration changes, or for regular health checks.

Comprehensive testing playbook for Local Brain Search memory improvements (Phases 1, 3, 4)

Free-form thinking mode - consider a topic through successive distinct knowledge-base perspectives (hub lenses, original frameworks, wildcard distant notes). One perspective pass per invocation, state persists across runs, emits [[DONE]] when the exploration weaves together. Loop-compatible by design.

Solve problems using knowledge base insights - extracts search terms, runs parallel KB queries, synthesizes advice grounded in your own frameworks

Systematic benchmarking framework for Local Brain Search memory system with LLM-as-judge scoring

Compare and selectively merge Brain directories across any two agent instances. Supports diff (see what's unique where), pull (import from another brain), and learn (bidirectional exchange).

Protocol for creating dated changelog files after significant agent sessions. Use when completing vault operations, discovery sessions, insight extraction, or any work requiring documentation.

An Electric Monk engine — two subagents believe fully committed positions on the user's behalf while the orchestrator performs structural contradiction analysis and synthesis. By outsourcing belief work to agents, the user operates from a belief-free position where they can analyze the structure of the contradiction rather than being inside either side. Use when the user wants to stress-test an idea, resolve a genuine tension, build a deeper mental model, or make a high-stakes decision where the tradeoffs are unclear. Works across any domain — technical architecture, product strategy, philosophy, personal decisions, risk analysis, policy, creative direction.

Advanced techniques for extracting insights, probing user thinking, and conducting cognitive interviews. Use when harvesting insights, exploring beliefs, clarifying concepts, or helping users articulate their unique perspectives.

Extract insights from external documents (research papers, books, articles). Spawns document-insight-extractor subagent. Requires session name.

Extract unique insights and perspectives from personal content (conversations, transcripts, notes). Spawns insight-extractor subagent.

Discover hidden connections and relationships between notes in the knowledge base

Stage, commit, and push changes to remote repository with approval gate

Standard format for capturing and documenting insights in the knowledge base. Use when harvesting insights, creating permanent notes from user perspectives, or documenting unique thinking patterns.

Manage domain-watch surveillance configs — seed new domains, review scan status and proposals, activate Watch Log proposals into thinking topics, pause domains, and update PIRs or watch queries