/**
* Operations Catalog for Gateway Toolhost
*
* Defines gateway-level operations (thought, read_thoughts, get_structure,
* cipher, deep_analysis) with their schemas, descriptions, and examples.
*/
export interface OperationDefinition {
name: string;
title: string;
description: string;
category: string;
inputSchema: any;
example?: any;
}
export const GATEWAY_OPERATIONS: OperationDefinition[] = [
{
name: "thought",
title: "Record Thought",
description:
"Record a structured reasoning thought in the active session. Server auto-assigns thoughtNumber and totalThoughts (SIL-102). Supports branching, revisions, and session metadata.",
category: "reasoning",
inputSchema: {
type: "object",
properties: {
thought: {
type: "string",
description: "The reasoning content to record",
},
nextThoughtNeeded: {
type: "boolean",
description: "Whether more thoughts are needed to complete the reasoning chain",
},
thoughtNumber: {
type: "number",
description: "Optional: thought sequence number (auto-assigned if omitted)",
},
totalThoughts: {
type: "number",
description: "Optional: estimated total thoughts (auto-assigned if omitted)",
},
branchId: {
type: "string",
description: "Branch name for parallel exploration (requires branchFromThought)",
},
branchFromThought: {
type: "number",
description: "Thought number to branch from (required when branchId is set)",
},
isRevision: {
type: "boolean",
description: "Whether this thought revises a previous thought",
},
revisesThought: {
type: "number",
description: "Thought number being revised (required when isRevision is true)",
},
verbose: {
type: "boolean",
description: "If true, return full response; if false/omitted, return minimal response (SIL-101)",
},
sessionTitle: {
type: "string",
description: "Set session title (first thought only)",
},
sessionTags: {
type: "array",
items: { type: "string" },
description: "Set session tags (first thought only)",
},
critique: {
type: "boolean",
description: "If true, include self-critique guidance",
},
},
required: ["thought", "nextThoughtNeeded"],
},
example: {
thought: "The architecture uses a gateway pattern to route operations...",
nextThoughtNeeded: true,
sessionTitle: "Architecture Analysis",
sessionTags: ["project:thoughtbox", "task:analysis"],
},
},
{
name: "read_thoughts",
title: "Read Thoughts",
description:
"Retrieve previous thoughts from the active or specified session. Supports multiple query modes: single thought by number, last N thoughts, range, or branch filtering.",
category: "reasoning",
inputSchema: {
type: "object",
properties: {
sessionId: {
type: "string",
description: "Session to read from (defaults to active session)",
},
thoughtNumber: {
type: "number",
description: "Get a single thought by number",
},
last: {
type: "number",
description: "Get the last N thoughts",
},
range: {
type: "array",
items: { type: "number" },
minItems: 2,
maxItems: 2,
description: "Get thoughts in range [start, end] (inclusive)",
},
branchId: {
type: "string",
description: "Get all thoughts from a named branch (standalone query mode — not combinable with thoughtNumber, last, or range)",
},
},
},
example: {
last: 5,
},
},
{
name: "get_structure",
title: "Get Reasoning Structure",
description:
"Return the topology of a reasoning session without thought content. Shows main chain length, branches, fork points, and revision pairs. Useful for understanding reasoning shape before reading specific thoughts.",
category: "reasoning",
inputSchema: {
type: "object",
properties: {
sessionId: {
type: "string",
description: "Session to analyze (defaults to active session)",
},
},
},
example: {},
},
{
name: "cipher",
title: "Load Cipher Notation",
description:
"Load the Thoughtbox cipher notation system for token-efficient reasoning. Must be called before thought operations. Advances stage from STAGE_1 to STAGE_2.",
category: "initialization",
inputSchema: {
type: "object",
properties: {},
},
example: {},
},
{
name: "deep_analysis",
title: "Deep Analysis",
description:
"Analyze a reasoning session for patterns, cognitive load, and decision points. Available analysis types: patterns, cognitive_load, decision_points, full.",
category: "analysis",
inputSchema: {
type: "object",
properties: {
sessionId: {
type: "string",
description: "Session to analyze",
},
analysisType: {
type: "string",
enum: ["patterns", "cognitive_load", "decision_points", "full"],
description: "Type of analysis to perform",
},
options: {
type: "object",
properties: {
includeTimeline: {
type: "boolean",
description: "Include session timeline in results",
},
compareWith: {
type: "array",
items: { type: "string" },
description: "Session IDs to compare with",
},
},
},
},
required: ["sessionId", "analysisType"],
},
example: {
sessionId: "abc-123-def-456",
analysisType: "full",
options: { includeTimeline: true },
},
},
];
/**
* Get operation definition by name
*/
export function getOperation(name: string): OperationDefinition | undefined {
return GATEWAY_OPERATIONS.find((op) => op.name === name);
}
/**
* Get all operation names
*/
export function getOperationNames(): string[] {
return GATEWAY_OPERATIONS.map((op) => op.name);
}
/**
* Get operations catalog as JSON resource
*/
export function getOperationsCatalog(): string {
return JSON.stringify(
{
version: "1.0.0",
operations: GATEWAY_OPERATIONS.map((op) => ({
name: op.name,
title: op.title,
description: op.description,
category: op.category,
inputs: op.inputSchema,
example: op.example,
})),
categories: [
{
name: "initialization",
description: "Stage progression and cipher loading",
},
{
name: "reasoning",
description: "Record, read, and explore thought chains",
},
{
name: "analysis",
description: "Deep analysis of reasoning sessions",
},
],
},
null,
2
);
}
