import { INTERLEAVED_THINKING_CONTENT } from "./contents/interleaved-thinking-content.js";
import {
interleavedGuide,
isValidMode,
type InterleavedMode,
} from "./contents/interleaved-template.js";
import {
PARALLEL_VERIFICATION_CONTENT,
getParallelVerificationContent,
parseParallelParams,
} from "./contents/parallel-verification.js";
import { SPEC_DESIGNER_CONTENT } from "./contents/spec-designer-content.js";
import { SPEC_VALIDATOR_CONTENT } from "./contents/spec-validator-content.js";
import { SPEC_ORCHESTRATOR_CONTENT } from "./contents/spec-orchestrator-content.js";
import { SPECIFICATION_SUITE_CONTENT } from "./contents/specification-suite-content.js";
export { LIST_MCP_ASSETS_PROMPT, getListMcpAssetsContent } from "./list-mcp-assets.js";
export { getParallelVerificationContent } from "./contents/parallel-verification.js";
/**
* MCP Prompt definition for interleaved thinking workflow
*/
export const INTERLEAVED_THINKING_PROMPT = {
name: "interleaved-thinking",
title: "interleaved-thinking-workflow",
description:
"Use this Thoughtbox server as a reasoning workspace to alternate between internal reasoning steps and external tool/action invocation. Enables structured multi-phase execution with tooling inventory, sufficiency assessment, strategy development, and execution.",
arguments: [
{
name: "task",
description: "The task to complete using interleaved thinking approach",
required: true,
},
{
name: "thoughts_limit",
description: "Maximum number of thoughts to use (default: 100)",
required: false,
},
{
name: "clear_folder",
description:
"Whether to clear the .interleaved-thinking folder after completion, keeping only final-answer.md (default: false)",
required: false,
},
],
};
/**
* MCP Prompt definition for parallel verification workflow
*/
export const PARALLEL_VERIFICATION_PROMPT = {
name: "parallel-verification",
title: "Parallel-Verification-Workflow",
description:
"Execute parallel hypothesis exploration using Thoughtbox. Pass your task/question - optional max_branches param can be included or will default to 3.",
arguments: [
{
name: "input",
description:
"Your task/question. May include 'max_branches:N' or just describe the task.",
required: true,
},
],
};
/**
* MCP Prompt definition for spec-designer workflow
*/
export const SPEC_DESIGNER_PROMPT = {
name: "spec-designer",
title: "Spec Designer",
description:
"Design and produce implementation specifications through structured cognitive loops. Creates specs from prompts using OODA loop building blocks.",
arguments: [
{
name: "prompt",
description: "The feature or system to design specifications for",
required: true,
},
{
name: "output_folder",
description: "Folder to output specs to (default: .specs/)",
required: false,
},
{
name: "depth",
description: "Spec depth: shallow, standard, or comprehensive (default: standard)",
required: false,
},
{
name: "max_specs",
description: "Maximum number of specs to generate (default: 5)",
required: false,
},
{
name: "plan_only",
description: "Only explore and plan, don't generate specs (default: false)",
required: false,
},
],
};
/**
* MCP Prompt definition for spec-validator workflow
*/
export const SPEC_VALIDATOR_PROMPT = {
name: "spec-validator",
title: "Spec Validator",
description:
"Systematically validate specification documents against current codebase and project architecture. Identifies gaps, contradictions, and feasibility issues.",
arguments: [
{
name: "spec_path",
description: "Path to spec file or folder to validate",
required: true,
},
{
name: "strict",
description: "Strict mode: any missing requirement is a failure (default: false)",
required: false,
},
{
name: "deep",
description: "Perform deeper semantic analysis and cross-file trace (default: false)",
required: false,
},
{
name: "report_only",
description: "Skip suggestion phase, only generate report (default: false)",
required: false,
},
],
};
/**
* MCP Prompt definition for spec-orchestrator workflow
*/
export const SPEC_ORCHESTRATOR_PROMPT = {
name: "spec-orchestrator",
title: "Spec Orchestrator",
description:
"Coordinate implementation of multiple specification documents from a folder. Manages dependencies, tracks progress, and prevents implementation spirals using Operations Research principles.",
arguments: [
{
name: "spec_folder",
description: "Folder containing specification documents to implement",
required: true,
},
{
name: "budget",
description: "Energy budget for implementation (default: 100)",
required: false,
},
{
name: "max_iterations",
description: "Maximum iterations per spec (default: 3)",
required: false,
},
{
name: "plan_only",
description: "Analyze without implementing (default: false)",
required: false,
},
],
};
/**
* MCP Prompt definition for specification-suite workflow
*/
export const SPECIFICATION_SUITE_PROMPT = {
name: "specification-suite",
title: "Specification Suite",
description:
"Chain the design → validate → orchestrate lifecycle into one command. Moves from blank prompt to implemented, validated specs.",
arguments: [
{
name: "prompt_or_spec_path",
description: "Prompt string OR existing spec folder/file path",
required: true,
},
{
name: "output_folder",
description: "Output folder for specs (default: .specs/)",
required: false,
},
{
name: "depth",
description: "Spec depth: shallow, standard, or comprehensive (default: standard)",
required: false,
},
{
name: "budget",
description: "Energy budget for orchestration (default: 100)",
required: false,
},
{
name: "plan_only",
description: "Stop after validation, don't implement (default: false)",
required: false,
},
{
name: "skip_design",
description: "Treat input as existing specs, skip design phase (default: false)",
required: false,
},
{
name: "skip_validation",
description: "Skip validation, jump from design to orchestration (default: false)",
required: false,
},
],
};
/**
* Loads the interleaved-thinking prompt content and substitutes variables
*/
export function getInterleavedThinkingContent(args: {
task: string;
thoughts_limit?: any;
clear_folder?: any;
}): string {
// Use bundled content (no file I/O)
let content = INTERLEAVED_THINKING_CONTENT;
// Apply defaults
const thoughtsLimit = args.thoughts_limit ?? 100;
const clearFolder = args.clear_folder ?? false;
// Substitute variables in the format shown in the markdown
// The markdown uses:
// TASK: $ARGUMENTS
// THOUGHTS_LIMIT: $ARGUMENTS (default: 100)
// CLEAR_FOLDER: $ARGUMENTS (default: false)
// Replace the variable declarations section
const variablesSection = `## Variables
TASK: $ARGUMENTS
THOUGHTS_LIMIT: $ARGUMENTS (default: 100)
CLEAR_FOLDER: $ARGUMENTS (default: false)`;
const substitutedVariables = `## Variables
TASK: ${args.task}
THOUGHTS_LIMIT: ${thoughtsLimit}
CLEAR_FOLDER: ${clearFolder}`;
content = content.replace(variablesSection, substitutedVariables);
return content;
}
/**
* Loads the spec-designer prompt content and substitutes variables
*/
export function getSpecDesignerContent(args: {
prompt: string;
output_folder?: string;
depth?: string;
max_specs?: string;
plan_only?: string;
}): string {
let content = SPEC_DESIGNER_CONTENT;
// Apply defaults
const outputFolder = args.output_folder ?? ".specs/";
const depth = args.depth ?? "standard";
const maxSpecs = args.max_specs ?? "5";
const planOnly = args.plan_only ?? "false";
// Substitute variables
const variablesSection = `## Variables
PROMPT: $ARGUMENTS
OUTPUT_FOLDER: $ARGUMENTS (default: .specs/)
DEPTH: $ARGUMENTS (default: standard)
MAX_SPECS: $ARGUMENTS (default: 5)
PLAN_ONLY: $ARGUMENTS (default: false)
CONFIDENCE_THRESHOLD: $ARGUMENTS (default: 0.85)`;
const substitutedVariables = `## Variables
PROMPT: ${args.prompt}
OUTPUT_FOLDER: ${outputFolder}
DEPTH: ${depth}
MAX_SPECS: ${maxSpecs}
PLAN_ONLY: ${planOnly}
CONFIDENCE_THRESHOLD: 0.85`;
content = content.replace(variablesSection, substitutedVariables);
return content;
}
/**
* Loads the spec-validator prompt content and substitutes variables
*/
export function getSpecValidatorContent(args: {
spec_path: string;
strict?: string;
deep?: string;
report_only?: string;
}): string {
let content = SPEC_VALIDATOR_CONTENT;
// Apply defaults
const strictMode = args.strict ?? "false";
const deepValidation = args.deep ?? "false";
const reportOnly = args.report_only ?? "false";
// Substitute variables
const variablesSection = `## Variables
SPEC_PATH: $ARGUMENTS
STRICT_MODE: $ARGUMENTS (default: false - if true, any missing requirement is a failure)
DEEP_VALIDATION: $ARGUMENTS (default: false - if true, performs deeper semantic analysis and cross-file trace)
REPORT_ONLY: $ARGUMENTS (default: false - skip suggestion phase)
RESUME: $ARGUMENTS (default: false)`;
const substitutedVariables = `## Variables
SPEC_PATH: ${args.spec_path}
STRICT_MODE: ${strictMode}
DEEP_VALIDATION: ${deepValidation}
REPORT_ONLY: ${reportOnly}
RESUME: false`;
content = content.replace(variablesSection, substitutedVariables);
return content;
}
/**
* Loads the spec-orchestrator prompt content and substitutes variables
*/
export function getSpecOrchestratorContent(args: {
spec_folder: string;
budget?: string;
max_iterations?: string;
plan_only?: string;
}): string {
let content = SPEC_ORCHESTRATOR_CONTENT;
// Apply defaults
const budget = args.budget ?? "100";
const maxIterations = args.max_iterations ?? "3";
const planOnly = args.plan_only ?? "false";
// Substitute variables
const variablesSection = `## Variables
SPEC_FOLDER: $ARGUMENTS
BUDGET: $ARGUMENTS (default: 100 energy units)
MAX_ITERATIONS: $ARGUMENTS (default: 3 per spec)
CONFIDENCE_THRESHOLD: $ARGUMENTS (default: 0.9)
WORKTREE_MODE: $ARGUMENTS (default: false)
PLAN_ONLY: $ARGUMENTS (default: false)`;
const substitutedVariables = `## Variables
SPEC_FOLDER: ${args.spec_folder}
BUDGET: ${budget}
MAX_ITERATIONS: ${maxIterations}
CONFIDENCE_THRESHOLD: 0.9
WORKTREE_MODE: false
PLAN_ONLY: ${planOnly}`;
content = content.replace(variablesSection, substitutedVariables);
return content;
}
/**
* Loads the specification-suite prompt content and substitutes variables
*/
export function getSpecificationSuiteContent(args: {
prompt_or_spec_path: string;
output_folder?: string;
depth?: string;
budget?: string;
plan_only?: string;
skip_design?: string;
skip_validation?: string;
}): string {
let content = SPECIFICATION_SUITE_CONTENT;
// Apply defaults
const outputFolder = args.output_folder ?? ".specs/";
const depth = args.depth ?? "standard";
const budget = args.budget ?? "100";
const planOnly = args.plan_only ?? "false";
const skipDesign = args.skip_design ?? "false";
const skipValidation = args.skip_validation ?? "false";
// Substitute variables
const variablesSection = `## Variables
PROMPT_OR_SPEC_PATH: $ARGUMENTS (prompt string OR existing spec folder/file)
OUTPUT_FOLDER: $ARGUMENTS (default: .specs/)
DEPTH: $ARGUMENTS (default: standard; passed to spec-designer)
MAX_SPECS: $ARGUMENTS (default: 5; passed to spec-designer)
STRICT_MODE: $ARGUMENTS (default: false; passed to spec-validator)
DEEP_VALIDATION: $ARGUMENTS (default: false; passed to spec-validator)
BUDGET: $ARGUMENTS (default: 100; passed to spec-orchestrator)
MAX_ITERATIONS: $ARGUMENTS (default: 3; passed to spec-orchestrator)
PLAN_ONLY: $ARGUMENTS (default: false; stop after validation)
RESUME: $ARGUMENTS (default: false; resume previous suite run)
SKIP_DESIGN: $ARGUMENTS (default: false; treat PROMPT_OR_SPEC_PATH as existing specs)
SKIP_VALIDATION: $ARGUMENTS (default: false; jump from design to orchestrate)`;
const substitutedVariables = `## Variables
PROMPT_OR_SPEC_PATH: ${args.prompt_or_spec_path}
OUTPUT_FOLDER: ${outputFolder}
DEPTH: ${depth}
MAX_SPECS: 5
STRICT_MODE: false
DEEP_VALIDATION: false
BUDGET: ${budget}
MAX_ITERATIONS: 3
PLAN_ONLY: ${planOnly}
RESUME: false
SKIP_DESIGN: ${skipDesign}
SKIP_VALIDATION: ${skipValidation}`;
content = content.replace(variablesSection, substitutedVariables);
return content;
}
/**
* Returns the resource template definition for interleaved thinking guides
* Used by the ListResourceTemplatesRequestSchema handler
*/
export function getInterleavedResourceTemplates() {
return {
resourceTemplates: [
{
uriTemplate: "thoughtbox://interleaved/{mode}",
name: "thoughtbox-interleaved-guides",
title: "Thoughtbox Interleaved Thinking Guides",
description:
"IRCoT-style interleaved reasoning guides centered on Thoughtbox as the canonical reasoning workspace. Mode parameter: research, analysis, or development.",
mimeType: "text/markdown",
annotations: {
audience: ["assistant"],
priority: 1.0,
},
},
],
};
}
/**
* Resolves a thoughtbox://interleaved/{mode} URI to its content
* Used by the ReadResourceRequestSchema handler
*
* @param uri - The resource URI to resolve
* @returns Resource content with metadata
* @throws Error if the URI is invalid or mode is not supported
*/
export function getInterleavedGuideForUri(uri: string) {
// Expected format: thoughtbox://interleaved/{mode}
const match = uri.match(/^thoughtbox:\/\/interleaved\/(.+)$/);
if (!match) {
throw new Error(
`Invalid interleaved guide URI: ${uri}. Expected format: thoughtbox://interleaved/{mode}`
);
}
const mode = match[1];
if (!isValidMode(mode)) {
throw new Error(
`Invalid mode: ${mode}. Must be one of: research, analysis, development`
);
}
return {
uri,
mimeType: "text/markdown",
text: interleavedGuide(mode),
};
}
