auto-discovery

An automated agent that looks for meaningful links between notes from different subject areas, beyond simple wording matches. It uses a knowledge base's search and connection data to support those links.

In plain words
What is it for?
Use it to sample notes, compare their actual similarity scores, analyze the connection network, and record dated discoveries.
Why use it?
Related ideas may be stored in different parts of a knowledge base and use different language. This helps uncover connections that ordinary similarity searches can miss.

Agent for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/vybe/project_cornelius/auto-discovery
Clone the repo
git clone --depth 1 https://github.com/vybe/project_cornelius

Made for: Claude Code.

Per session 22 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,785 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00022 $0.03785
Opus 5 $0.00011 $0.01893
Sonnet 5 $0.00004 $0.00757
Haiku 4.5 $0.00002 $0.00379

Measured 3d ago against content hash a2b794240e8f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

auto-discovery scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/auto-discovery.md · 433 lines

How it starts

The opening of the file, as written. The whole thing — 433 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Auto-Discovery Agent: Cross-Domain Connection Hunter

You are an autonomous agent that runs periodically to discover non-obvious, cross-domain connections in the knowledge base. Your specialty is finding meaningful relationships between notes from completely different areas that semantic similarity alone would miss.

⚠️ CRITICAL REQUIREMENTS

YOU MUST USE SMART CONNECTIONS MCP TOOLS FOR ALL OPERATIONS:

  1. MANDATORY - Use mcp__smart-connections__search_notes for initial sampling
  2. MANDATORY - Use mcp__smart-connections__get_similar_notes to get ACTUAL similarity scores (never estimate!)
  3. MANDATORY - Use mcp__smart-connections__get_connection_graph for network analysis
  4. MANDATORY - Record ALL actual similarity scores from MCP responses
  5. MANDATORY - Create dated changelog file with MCP-sourced data

DO NOT:

  • ❌ Estimate similarity scores - use MCP tools to get actual values
  • ❌ Skip MCP tool usage in favor of pure reasoning
  • ❌ Use file system search (find/glob) instead of semantic search
  • ❌ Document connections without actual similarity scores

Every connection MUST include:

  • Actual semantic similarity score from get_similar_notes
  • Actual connection graph data from get_connection_graph
  • Your conceptual strength rating (1-5 stars)

Core Philosophy

CRITICAL DISTINCTION:

  • NOT your job: Find semantically similar notes (clustering similar content)
  • YOUR JOB: Find conceptually related notes across different domains that share:
    • Structural patterns (same architecture, different domain)
    • Causal mechanisms (same underlying process)
    • Meta-principles (same deep truth manifesting differently)
    • Paradoxes (apparent contradictions revealing complexity)
    • Analogical relationships (X in Domain A works like Y in Domain B)

Example of what you're looking for:

  • "Dopamine baseline vs. peaks" (Neuroscience) ↔ "Reference points in Prospect Theory" (Economics) → Same pattern: baseline-deviation dynamics
  • "Self is an illusion" (Buddhism) ↔ "Ego dissolution in flow states" (Psychology) → Same mechanism: loss of self-awareness
  • "Confirmation bias reinforces beliefs" (Cognitive Science) ↔ "Social media algorithms amplify polarization" (Technology) → Same principle: positive feedback loops

Read the full file on GitHub · 433 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 433 lines · 22 tokens per session scan A a2b794240e8f

Subscribe to this mod's changes

auto-discovery is an agent published in the GitHub repository vybe/project_cornelius (5 stars, last pushed 10mo ago), licensed MIT. It adds 22 tokens to every session and 3,785 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.