Tool Reference

⍌

Signal

Activation, flow, propagation. How information spreads through the graph.

4 tools

⍌ activate — Spreading activation query

Find related nodes by propagating signal through the graph. Starts from a query, activates matching nodes, then spreads energy along weighted edges. Returns ranked results by activation strength. The core discovery mechanism of the connectome.

Parameters

Name	Type	Description
query	string	What to search for. Can be a concept, module name, or natural language question.
depth	1-5	How many hops to propagate. Higher = broader but noisier results.
threshold	0.0-1.0	Minimum activation energy to include in results. Higher = more precise.
agent_id	string	Calling agent identifier for multi-agent coordination.

Example

// Find everything related to authentication
m1nd.activate({
  "query": "authentication middleware",
  "depth": 3,
  "threshold": 0.4,
  "agent_id": "jimi"
})

Exploration Research Impact scoping Onboarding

⍌ ingest — Load codebase into graph

Parse a codebase or JSON data and build the connectome graph. Extracts modules, functions, types, and files as nodes. Discovers imports, calls, contains, and co-change relationships as edges. Supports full rebuild or incremental updates.

Parameters

Name	Type	Description
source	path	Path to the codebase directory or JSON file to ingest.
mode	full \| incremental	Full rebuilds the entire graph. Incremental updates only changed files.
agent_id	string	Calling agent identifier.

Example

// Ingest a Rust project
m1nd.ingest({
  "source": "./my-project",
  "mode": "full",
  "agent_id": "jimi"
})

Session start After code changes Codebase onboarding

⍌ learn — Hebbian feedback

Strengthen or weaken graph connections based on usage feedback. When you use activate and the results are helpful, call learn with "correct". If results were off, use "wrong". The graph adapts over time, improving result quality through Hebbian plasticity.

Parameters

Name	Type	Description
query	string	The original query that produced results.
feedback	correct \| wrong \| partial	Was the result useful? Adjusts edge weights accordingly.
agent_id	string	Calling agent identifier.

Example

// The activation results were on point
m1nd.learn({
  "query": "authentication middleware",
  "feedback": "correct",
  "agent_id": "jimi"
})

After every activate Graph tuning Quality improvement

⍌ warmup — Prime graph for a task

Pre-activate a relevant subgraph before starting focused work. Primes the spreading activation network so subsequent queries in the same domain return faster, more relevant results. Think of it as "warming up the neural pathways."

Parameters

Name	Type	Description
context	string	Description of the task you are about to work on.
agent_id	string	Calling agent identifier.

Example

// About to refactor the WebSocket layer
m1nd.warmup({
  "context": "Refactoring WebSocket reconnection and pub/sub fan-out",
  "agent_id": "jimi"
})

Before focused work Task preparation Domain priming

⍐

Path

Tracing, dependency chains, temporal ordering, directed search.

4 tools

⍐ why — Explain path between two nodes

Trace the dependency chain between two nodes in the graph. Answers "Why does A relate to B?" by finding the shortest weighted path and explaining each hop. Critical for understanding hidden coupling and non-obvious dependencies.

Parameters

Name	Type	Description
from	string	Source node (module, function, or file).
to	string	Target node.
max_hops	number	Maximum path length to search.
agent_id	string	Calling agent identifier.

Example

m1nd.why({
  "from": "auth_middleware",
  "to": "billing_service",
  "max_hops": 5,
  "agent_id": "jimi"
})

Understanding dependencies Tracing influence Architecture analysis

⍐ trace — Full dependency trace from a node

Follow all connections from a node in a given direction. Upstream shows what depends on it. Downstream shows what it depends on. Both gives the complete dependency web.

Parameters

Name	Type	Description
node	string	Starting node for the trace.
direction	upstream \| downstream \| both	Which direction to follow edges.
depth	number	How deep to trace.
agent_id	string	Calling agent identifier.

Example

m1nd.trace({
  "node": "database_pool",
  "direction": "upstream",
  "depth": 4,
  "agent_id": "jimi"
})

Impact analysis Refactoring prep Architecture mapping

⍐ timeline — Temporal ordering of changes

View the temporal history of a node. Shows when it was created, modified, and what co-changed alongside it. Powered by the TemporalEngine's causal chain analysis.

Parameters

Name	Type	Description
node	string	Node to get the timeline for.
range	string	Time range filter (e.g. "7d", "30d", "all").
agent_id	string	Calling agent identifier.

Example

m1nd.timeline({
  "node": "config.py",
  "range": "7d",
  "agent_id": "jimi"
})

Change archaeology Debugging regressions History analysis

⍐ seek — Directed search with semantic ranking

A more targeted search than activate. Uses semantic ranking within a constrained scope. Best for when you know roughly where to look but need the graph to rank results intelligently.

Parameters

Name	Type	Description
query	string	What to search for.
scope	string	Constrain search to a subgraph region.
limit	number	Maximum results to return.
agent_id	string	Calling agent identifier.

Example

m1nd.seek({
  "query": "error handling patterns",
  "scope": "backend/",
  "limit": 10,
  "agent_id": "jimi"
})

Focused search Code review Pattern discovery

⍂

Structure

Gaps, holes, duplicates, divergence. The shape of what is missing.

4 tools

⍂ missing — Find structural holes

Identify gaps in the graph structure. Uses topology analysis (structural holes, betweenness centrality) to find places where connections should exist but do not. The signature capability -- "find what's missing" is what separates a graph engine from a search engine.

Parameters

Name	Type	Description
context	string	What domain to analyze for gaps.
scope	string	Constrain analysis to a subgraph.
agent_id	string	Calling agent identifier.

Example

m1nd.missing({
  "context": "error handling",
  "scope": "backend/api/",
  "agent_id": "jimi"
})

Gap analysis Pre-spec validation Architecture review Test coverage

⍂ fingerprint — Duplicate and equivalence detection

Detect structurally similar or duplicate modules. Compares graph topology (not just code text) to find modules that serve the same purpose. Critical for consolidation and audit.

Parameters

Name	Type	Description
nodes	string[]	Nodes to fingerprint and compare.
threshold	0.0-1.0	Similarity threshold for matches.
agent_id	string	Calling agent identifier.

Example

m1nd.fingerprint({
  "nodes": ["auth_service", "legacy_auth"],
  "threshold": 0.7,
  "agent_id": "jimi"
})

Audit Consolidation Post-build review

⍂ scan — Structural scan of graph region

Scan a region of the graph and report its structural properties: density, clustering coefficient, central nodes, isolated nodes, and community boundaries.

Parameters

Name	Type	Description
scope	string	Region to scan.
depth	number	Depth of structural analysis.
agent_id	string	Calling agent identifier.

Example

m1nd.scan({
  "scope": "frontend/components/",
  "depth": 3,
  "agent_id": "jimi"
})

Architecture mapping Health checks Codebase exploration

⍂ diverge — Find divergence points

Compare two subgraphs to find where they diverge structurally. Useful for comparing feature branches, old vs new implementations, or cross-service boundaries.

Parameters

Name	Type	Description
a	string	First subgraph scope.
b	string	Second subgraph scope.
agent_id	string	Calling agent identifier.

Example

m1nd.diverge({
  "a": "services/auth-v1",
  "b": "services/auth-v2",
  "agent_id": "jimi"
})

Migration planning Branch comparison Drift detection

𝔻

Dimension

Blast radius, prediction, what-if simulation, differential analysis.

5 tools

𝔻 impact — Blast radius analysis

Determine what would be affected by changing a node. Propagates impact through dependency edges to show the "blast radius" of a modification. Essential before any refactoring.

Parameters

Name	Type	Description
node	string	The node you plan to modify.
depth	number	How far to propagate impact.
agent_id	string	Calling agent identifier.

Example

m1nd.impact({
  "node": "database_schema",
  "depth": 4,
  "agent_id": "jimi"
})

Pre-refactor analysis Scope validation Risk assessment

𝔻 predict — Co-change prediction

Given a set of changed nodes, predict what else needs to change. Uses co-change history from the TemporalEngine to identify files that historically change together.

Parameters

Name	Type	Description
changed_nodes	string[]	Nodes that have been or will be modified.
agent_id	string	Calling agent identifier.

Example

m1nd.predict({
  "changed_nodes": ["user_model", "user_routes"],
  "agent_id": "jimi"
})

After modifying code Before committing Completeness check

𝔻 counterfactual — Simulate node removal

Simulate removing nodes from the graph and see what breaks. The CounterfactualEngine creates a shadow graph, removes specified nodes, and reports disconnections, orphans, and broken paths.

Parameters

Name	Type	Description
remove_nodes	string[]	Nodes to simulate removing.
agent_id	string	Calling agent identifier.

Example

m1nd.counterfactual({
  "remove_nodes": ["legacy_adapter", "compat_layer"],
  "agent_id": "jimi"
})

Before deleting code Migration planning Deprecation analysis

𝔻 differential — Compare two graph states

Compare the graph at two different points in time or across two snapshots. Shows nodes added, removed, and edges that changed weight. Useful for understanding what happened between sessions.

Parameters

Name	Type	Description
baseline	string	Baseline snapshot or timestamp.
current	string	Current snapshot or "now".
agent_id	string	Calling agent identifier.

Example

m1nd.differential({
  "baseline": "2026-03-10",
  "current": "now",
  "agent_id": "jimi"
})

Session review Changelog generation Progress tracking

𝔻 hypothesize — What-if scenario modeling

Model hypothetical changes without actually modifying the graph. Describe a scenario and the engine simulates its effects on the connectome topology.

Parameters

Name	Type	Description
hypothesis	string	Description of the hypothetical change.
agent_id	string	Calling agent identifier.

Example

m1nd.hypothesize({
  "hypothesis": "What if we split the monolithic API into 3 microservices?",
  "agent_id": "jimi"
})

Architecture planning Design decisions Risk modeling

⍂

State

Health, drift, locks, trails, validation. Managing graph lifecycle and session continuity.

12 tools

⍂ health — Server diagnostics

Check the health of the m1nd server. Reports node count, edge count, memory usage, uptime, query count, and persistence status.

Parameters

Name	Type	Description
agent_id	string	Calling agent identifier.

Example

m1nd.health({ "agent_id": "jimi" })

MonitoringDebuggingSession start

⍂ drift — Weight changes since baseline

Report what has changed in the graph since a given point. Essential for session recovery -- call at the start of every new session to understand what happened since your last context.

Parameters

Name	Type	Description
since	string	Timestamp or label (e.g. "last_session", "2h").
agent_id	string	Calling agent identifier.

Example

m1nd.drift({ "since": "last_session", "agent_id": "jimi" })

Session startContext recovery

⍂ validate.plan — Validate implementation plan against graph

Check whether a proposed implementation plan is consistent with the current graph state. Identifies missing dependencies, circular references, and unreachable goals.

Parameters

Name	Type	Description
plan	string	The plan to validate (JSON or description).
agent_id	string	Calling agent identifier.

Example

m1nd.validate_plan({
  "plan": "Refactor auth: split into auth-core and auth-oauth",
  "agent_id": "jimi"
})

Pre-build validationPlan review

⍂ lock.create — Lock a subgraph region

Create a snapshot lock on a subgraph region. While locked, you can track changes (lock.diff), rebase to current state (lock.rebase), or get notified of modifications (lock.watch).

Parameters

Name	Type	Description
scope	string	Region to lock.
agent_id	string	Calling agent identifier.

Example

m1nd.lock_create({ "scope": "backend/auth/", "agent_id": "jimi" })

⍂lock.watch— Watch for changes in locked region

Monitor a locked region for changes made by other agents. Choose a strategy: alert on any change, or only structural changes.

Parameters

Name	Type	Description
lock_id	string	Lock identifier from lock.create.
strategy	string	Watch strategy (any, structural).
agent_id	string	Calling agent identifier.

⍂lock.diff— Diff changes since lock baseline

Compare the current graph state against the locked baseline. Shows what changed while you held the lock.

Parameters

Name	Type	Description
lock_id	string	Lock identifier.
agent_id	string	Calling agent identifier.

⍂lock.rebase— Rebase lock to current state

Update the lock baseline to the current graph state. Use after reviewing lock.diff to accept observed changes.

Parameters

Name	Type	Description
lock_id	string	Lock identifier.
agent_id	string	Calling agent identifier.

⍂lock.release— Release a graph lock

Release a previously created lock, cleaning up the baseline snapshot.

Parameters

Name	Type	Description
lock_id	string	Lock identifier.
agent_id	string	Calling agent identifier.

⍂trail.save— Save exploration trail

Save the current exploration trail (sequence of queries and activations) as a named bookmark. Resume later with trail.resume.

Parameters

Name	Type	Description
name	string	Name for this trail.
agent_id	string	Calling agent identifier.

⍂trail.resume— Resume saved trail

Resume a previously saved trail, restoring the activation state and exploration context.

Parameters

Name	Type	Description
trail_id	string	Trail identifier from trail.save.
agent_id	string	Calling agent identifier.

⍂trail.merge— Merge two trails

Combine two exploration trails into one, merging their activation histories and context.

Parameters

Name	Type	Description
trail_a	string	First trail identifier.
trail_b	string	Second trail identifier.
agent_id	string	Calling agent identifier.

⍂trail.list— List all trails

List all saved exploration trails with their names, creation times, and activation summaries.

Parameters

Name	Type	Description
agent_id	string	Calling agent identifier.

⟁

Connection

Perspectives, federation, resonance. How parts relate across boundaries.

14 tools

⟁ resonate — Standing wave / harmonic analysis

Find resonance patterns in the graph. The ResonanceEngine sends signals from seed nodes and detects standing wave patterns -- nodes that sustain high activation from multiple sources simultaneously. Reveals deep structural relationships invisible to regular search.

Parameters

Name	Type	Description
seeds	string[]	Seed nodes to emit from.
harmonics	number	Number of harmonic frequencies to analyze.
agent_id	string	Calling agent identifier.

Example

m1nd.resonate({
  "seeds": ["auth", "billing", "notifications"],
  "harmonics": 3,
  "agent_id": "jimi"
})

Deep structural analysis Pattern discovery Cross-domain relationships

⟁ federate — Cross-graph federation query

Query across multiple graph instances simultaneously. Useful for multi-repo projects or comparing different codebases. Returns unified results with source attribution.

Parameters

Name	Type	Description
query	string	What to search for across graphs.
graphs	string[]	Graph identifiers to query.
agent_id	string	Calling agent identifier.

Example

m1nd.federate({
  "query": "database migrations",
  "graphs": ["backend", "data-pipeline"],
  "agent_id": "jimi"
})

Multi-repo search Cross-project analysis

⟁ perspective.start — Start a perspective session

Begin an interactive graph exploration session. A perspective is a stateful walk through the graph -- you focus on a node, see its connections, navigate along edges, and build understanding through movement rather than querying.

Parameters

Name	Type	Description
focus	string	Starting node for the perspective.
agent_id	string	Calling agent identifier.

Example

m1nd.perspective_start({
  "focus": "main.py",
  "agent_id": "jimi"
})

⟁perspective.routes— Get navigation routes from focus

List all navigable routes from the current perspective focus. Shows connected nodes grouped by edge type.

Parameters

Name	Type	Description
perspective_id	string	Active perspective identifier.
agent_id	string	Calling agent identifier.

⟁perspective.inspect— Deep inspect current node

Get full details about the node at your current perspective focus. Properties, metadata, connection count, weight distribution, and community membership.

Parameters

Name	Type	Description
perspective_id	string	Active perspective identifier.
agent_id	string	Calling agent identifier.

⟁perspective.peek— Preview without navigating

Preview a connected node without changing your perspective focus. Like peeking around a corner.

Parameters

Name	Type	Description
perspective_id	string	Active perspective identifier.
node	string	Node to peek at.
agent_id	string	Calling agent identifier.

⟁perspective.follow— Navigate to a connected node

Move your perspective focus along an edge to a connected node. Your navigation history is preserved for backtracking.

Parameters

Name	Type	Description
perspective_id	string	Active perspective identifier.
route	string	Route/edge to follow.
agent_id	string	Calling agent identifier.

⟁perspective.suggest— AI-suggested next moves

Get intelligent suggestions for where to navigate next based on your exploration history and current position. The engine analyzes patterns in your traversal to suggest high-value next steps.

Parameters

Name	Type	Description
perspective_id	string	Active perspective identifier.
agent_id	string	Calling agent identifier.

⟁perspective.affinity— Find high-affinity nodes

Find nodes with the strongest connection to your current focus. Unlike routes (which show direct connections), affinity considers multi-hop weighted paths.

Parameters

Name	Type	Description
perspective_id	string	Active perspective identifier.
agent_id	string	Calling agent identifier.

⟁perspective.branch— Branch into parallel exploration

Fork your current perspective into a new branch. Both branches share history up to this point but diverge from here. Compare later with perspective.compare.

Parameters

Name	Type	Description
perspective_id	string	Active perspective identifier.
agent_id	string	Calling agent identifier.

⟁perspective.back— Navigate back in history

Go back to the previous node in your perspective history. Like the back button in a browser.

Parameters

Name	Type	Description
perspective_id	string	Active perspective identifier.
agent_id	string	Calling agent identifier.

⟁perspective.compare— Compare two perspectives

Compare the paths and discoveries of two perspective sessions. Shows overlapping nodes, divergence points, and unique findings.

Parameters

Name	Type	Description
perspective_a	string	First perspective.
perspective_b	string	Second perspective.
agent_id	string	Calling agent identifier.

⟁perspective.list— List active perspectives

List all active perspective sessions with their current focus, history length, and creation time.

Parameters

Name	Type	Description
agent_id	string	Calling agent identifier.

⟁perspective.close— Close a perspective session

Close and clean up a perspective session. Consider saving it as a trail first with trail.save if you want to resume later.

Parameters

Name	Type	Description
perspective_id	string	Perspective to close.
agent_id	string	Calling agent identifier.