@ryuenn3123/agentic-senior-core 2.0.7 → 2.0.9

package/.agent-context/prompts/review-code.md

@@ -22,11 +22,11 @@ For EVERY violation found:
 
 Output format:
 ## PR REVIEW RESULTS
-✅ [Item] — Passes
-❌ [Item] — FAILS (with Reasoning Chain)
+PASS [Item]
+FAIL [Item] (with Reasoning Chain)
 
 ## SECURITY AUDIT RESULTS
-🔴/🟠/🟡/🟢 [Finding] — severity + fix
+CRITICAL/HIGH/MEDIUM/LOW [Finding] — severity + fix
 
 VERDICT: PASS / FAIL
 ```

package/.agent-context/review-checklists/pr-checklist.md

@@ -13,13 +13,13 @@ Output format:
 ## PR REVIEW RESULTS
 ━━━━━━━━━━━━━━━━━━━
 
-✅ [Item] — Passes
-❌ [Item] — FAILS
-   📌 Rule: [rule file + section]
-   ❌ Problem: [specific issue found]
-   ✅ Fix: [what to change]
+PASS [Item]
+FAIL [Item]
+   Rule: [rule file + section]
+   Problem: [specific issue found]
+   Fix: [what to change]
 
-VERDICT: PASS ✅ / FAIL ❌ (X/Y items passed)
+VERDICT: PASS / FAIL (X/Y items passed)
 ```
 
 ---
@@ -95,3 +95,8 @@ VERDICT: PASS ✅ / FAIL ❌ (X/Y items passed)
 - [ ] Complex business logic has comments explaining WHY
 - [ ] Public functions/methods have JSDoc/docstrings
 - [ ] README updated if new setup steps required
+- [ ] No emoji in formal documentation or review summaries
+- [ ] Documentation uses plain English and avoids AI cliches
+- [ ] Performance/quality claims include source and timestamp
+- [ ] Acronyms are expanded on first use
+- [ ] Facts and assumptions are explicitly separated

package/.agent-context/rules/api-docs.md

@@ -25,6 +25,40 @@ REQUIRED: Documentation MUST be updated in the SAME commit as the endpoint chang
 
 ---
 
+## Human Writing Standard (Mandatory)
+
+This standard applies to API docs, README updates, release notes, and technical explanations.
+
+### Style Baseline
+1. Write for native English speakers.
+2. Target an 8th-grade reading level.
+3. Use clear, direct, plain language.
+4. Keep sentence rhythm natural with short and medium sentences.
+5. Sound confident, practical, and conversational.
+
+### Required Behavior
+1. Explain decisions the way a competent coworker would explain them out loud.
+2. Cut unnecessary words and remove filler.
+3. Use concrete verbs and everyday phrasing.
+4. Rewrite and reorder content when flow is weak.
+
+### Hard Bans
+1. No emoji in formal artifacts.
+2. No AI cliches: delve, leverage, robust, utilize, seamless.
+3. No inflated, academic, or performative language.
+4. No padding, hedging, or redundant phrasing.
+
+### Critical Controls
+1. Any claim about quality, performance, or reliability must include a measurable source and timestamp.
+2. Expand acronyms on first use, then use terms consistently.
+3. Separate facts from assumptions explicitly.
+4. End major explanations with a clear next action.
+
+### Final Check
+Read the text out loud before shipping. If it sounds robotic, rewrite it.
+
+---
+
 ## Documentation Format: OpenAPI 3.1 (Non-Negotiable)
 
 All APIs produce an OpenAPI 3.1 specification. Not 3.0, not proprietary formats, not "we'll add Swagger later."

package/.agent-context/skills/review-quality.md

@@ -19,6 +19,9 @@ Turn code review, planning, and benchmark output into explicit quality decisions
 - Benchmark deltas are explicit
 - Security findings stop release when critical
 - Release readiness verdict includes blockers, owner, and due date
+- Formal review artifacts use plain, human-first language and avoid AI cliches
+- No emoji in formal review output, release notes, and governance summaries
+- Claims include measurable evidence with source and timestamp
 
 ## Evidence
 - PR review report

package/.cursorrules

@@ -1,6 +1,6 @@
 # AGENTIC-SENIOR-CORE DYNAMIC GOVERNANCE RULESET
 
-Generated by Agentic-Senior-Core CLI v2.0.7
+Generated by Agentic-Senior-Core CLI v2.0.9
 Timestamp: 2026-04-08T14:58:53.570Z
 Selected profile: beginner
 Selected policy file: .agent-context/policies/llm-judge-threshold.json

package/.gemini/instructions.md

@@ -2,7 +2,7 @@
 
 Adapter Mode: thin
 Adapter Source: .instructions.md
-Canonical Snapshot SHA256: 361ac86172268a2f20d4f7a1e407c7dd122f1b0c265d085e872254718b078d22
+Canonical Snapshot SHA256: 08994f6e228b32d415dcc2024f31f0a076119a3cf87a4cd2fd2e78e86e5fbd3e
 
 Canonical policy source: [.instructions.md](../.instructions.md).
 

package/.github/copilot-instructions.md

@@ -2,7 +2,7 @@
 
 Adapter Mode: thin
 Adapter Source: .instructions.md
-Canonical Snapshot SHA256: 361ac86172268a2f20d4f7a1e407c7dd122f1b0c265d085e872254718b078d22
+Canonical Snapshot SHA256: 08994f6e228b32d415dcc2024f31f0a076119a3cf87a4cd2fd2e78e86e5fbd3e
 
 The canonical policy source for this repository is [.instructions.md](../.instructions.md).
 
@@ -18,4 +18,4 @@ The canonical policy source for this repository is [.instructions.md](../.instru
 
 ## Completion Gate
 
-Run [.agent-context/review-checklists/pr-checklist.md](../.agent-context/review-checklists/pr-checklist.md) before declaring work complete.
+Run [.agent-context/review-checklists/pr-checklist.md](../.agent-context/review-checklists/pr-checklist.md) before declaring work complete.

package/.windsurfrules

@@ -1,6 +1,6 @@
 # AGENTIC-SENIOR-CORE DYNAMIC GOVERNANCE RULESET
 
-Generated by Agentic-Senior-Core CLI v2.0.7
+Generated by Agentic-Senior-Core CLI v2.0.9
 Timestamp: 2026-04-08T14:58:53.570Z
 Selected profile: beginner
 Selected policy file: .agent-context/policies/llm-judge-threshold.json

package/AGENTS.md

@@ -2,7 +2,7 @@
 
 Adapter Mode: thin
 Adapter Source: .instructions.md
-Canonical Snapshot SHA256: 361ac86172268a2f20d4f7a1e407c7dd122f1b0c265d085e872254718b078d22
+Canonical Snapshot SHA256: 08994f6e228b32d415dcc2024f31f0a076119a3cf87a4cd2fd2e78e86e5fbd3e
 
 This file is an adapter entrypoint for agent discovery.
 The canonical policy source is [.instructions.md](.instructions.md).

package/README.md

@@ -119,6 +119,26 @@ npx @ryuenn3123/agentic-senior-core init --profile-pack startup
 
 The CLI is smart. It auto-detects your current development stack, helps you build a governance profile (select from `beginner`, `balanced`, or `strict`), and writes the compiled rules straight to your root automatically!
 
+Important behavior:
+- `init` does not copy repository workflows from this project into your target repository.
+- MCP server registration and trust/start are managed manually in IDE settings.
+- If you want the MCP workspace scaffold, run `init` with `--mcp-template` (creates `.vscode/mcp.json`).
+
+### MCP Setup in VS Code (No File Picker)
+
+If you are looking for a file picker in the MCP UI, that is expected because VS Code uses MCP server registration, not generic JSON file import.
+
+1. Generate workspace MCP config:
+
+```bash
+npx @ryuenn3123/agentic-senior-core init --mcp-template
+```
+
+2. Open Command Palette and run `MCP: Open Workspace Folder Configuration`.
+3. Confirm the file is `.vscode/mcp.json` with server `agentic-senior-core`.
+4. The generated server command is `npx -y @ryuenn3123/agentic-senior-core mcp`.
+5. Open Chat Customizations > MCP Servers, then trust/start the server.
+
 If you are totally new to concepts like blueprints and guardrails, no problem — just run:
 ```bash
 npx @ryuenn3123/agentic-senior-core init --newbie
@@ -265,7 +285,8 @@ Our documentation has shifted into dedicated tracks to keep this README light:
 - **Automated Guardrails:** CI blueprints include LLM-as-a-Judge flow using `pr-checklist.md`.
 - **Pre-Publish Safety:** Built-in forbidden content checks detect hardcoded secrets and stray debugger artifacts before hitting the NPM registry.
 - **Machine-Readable CI Output:** LLM Judge emits `JSON_REPORT` payloads and writes `.agent-context/state/llm-judge-report.json` for PR/MR annotation tooling.
-- **MCP Self-Healing Loop:** `mcp.json` defines diagnostics + fix proposal workflow when lint/CI fails.
+- **MCP Runtime Server:** `scripts/mcp-server.mjs` exposes validate/test/release checks as MCP tools.
+- **MCP Registration Model:** IDE MCP server registration is manual; workspace config lives in `.vscode/mcp.json` (`--mcp-template`).
 
 ---
 
@@ -276,7 +297,9 @@ Our documentation has shifted into dedicated tracks to keep this README light:
 ├── .cursorrules                    # Dynamic compiled governance entry point
 ├── .windsurfrules                  # Dynamic compiled governance entry point
 ├── .agent-override.md              # Team-specific exceptions (scoped + expiry)
-├── mcp.json                        # MCP self-healing workflow config
+├── mcp.json                        # Governance metadata and knowledge-layer contract
+├── .vscode/
+│   └── mcp.json                    # VS Code MCP workspace server configuration
 ├── AGENTS.md                       # Universal agent discovery
 ├── .github/copilot-instructions.md # GitHub Copilot entry point
 ├── .gemini/instructions.md         # Antigravity / Gemini entry point
@@ -294,6 +317,7 @@ Our documentation has shifted into dedicated tracks to keep this README light:
 ├── scripts/
 │   ├── validate.mjs                # Repository validator
 │   ├── llm-judge.mjs               # LLM-as-a-Judge CI gate
+│   ├── mcp-server.mjs              # Local MCP stdio server (validate/test/release tools)
 │   ├── init-project.sh             # GitHub bootstrap script (Linux/macOS)
 │   └── init-project.ps1            # GitHub bootstrap script (Windows)
 ├── docs/

package/bin/agentic-senior-core.js

@@ -11,6 +11,7 @@ import { CLI_VERSION } from '../lib/cli/constants.mjs';
 import { printUsage } from '../lib/cli/utils.mjs';
 import { runLaunchCommand } from '../lib/cli/commands/launch.mjs';
 import { runRollbackCommand } from '../lib/cli/commands/rollback.mjs';
+import { runMcpServerCommand } from '../lib/cli/commands/mcp.mjs';
 import { runOptimizeCommand, parseOptimizeArguments } from '../lib/cli/commands/optimize.mjs';
 import { runInitCommand, parseInitArguments } from '../lib/cli/commands/init.mjs';
 import { runUpgradeCommand, parseUpgradeArguments } from '../lib/cli/commands/upgrade.mjs';
@@ -63,6 +64,11 @@ async function main() {
     return;
   }
 
+  if (commandArgument === 'mcp') {
+    await runMcpServerCommand();
+    return;
+  }
+
   console.error(`Unknown command: ${commandArgument}`);
   printUsage();
   exit(1);

package/lib/cli/commands/init.mjs

@@ -57,6 +57,7 @@ export function parseInitArguments(commandArguments) {
     newbie: false,
     tokenOptimize: true,
     tokenAgent: 'copilot',
+    includeMcpTemplate: false,
   };
 
   for (let argumentIndex = 0; argumentIndex < commandArguments.length; argumentIndex++) {
@@ -160,6 +161,11 @@ export function parseInitArguments(commandArguments) {
       continue;
     }
 
+    if (currentArgument === '--mcp-template') {
+      parsedInitOptions.includeMcpTemplate = true;
+      continue;
+    }
+
     throw new Error(`Unknown option: ${currentArgument}`);
   }
 
@@ -177,6 +183,7 @@ export async function runInitCommand(targetDirectoryArgument, initOptions = {})
   const isTokenOptimizationEnabled = typeof initOptions.tokenOptimize === 'boolean'
     ? initOptions.tokenOptimize
     : true;
+  const shouldIncludeMcpTemplate = initOptions.includeMcpTemplate === true;
   const selectedTokenAgentName = normalizeAgentName(initOptions.tokenAgent || 'copilot');
 
   if (resolvedTargetDirectoryPath.toLowerCase() === 'c:\\windows' || resolvedTargetDirectoryPath.toLowerCase() === 'c:\\windows\\system32') {
@@ -348,7 +355,9 @@ export async function runInitCommand(targetDirectoryArgument, initOptions = {})
         ? selectedProfile.defaultCi
         : await askYesNo('Enable CI/CD guardrails and the LLM Judge policy?', userInterface, selectedProfile.defaultCi);
 
-    await copyGovernanceAssetsToTarget(resolvedTargetDirectoryPath);
+    await copyGovernanceAssetsToTarget(resolvedTargetDirectoryPath, {
+      includeMcpTemplate: shouldIncludeMcpTemplate,
+    });
 
     if (isTokenOptimizationEnabled) {
       const detectedExternalProxy = detectRtkBinary();
@@ -419,6 +428,8 @@ export async function runInitCommand(targetDirectoryArgument, initOptions = {})
     console.log(`- Blocking severities: ${formatBlockingSeverities(selectedProfile.blockingSeverities)}`);
     console.log(`- Setup time: ${formatDuration(setupDurationMs)}`);
     console.log('- Generated files: .cursorrules, .windsurfrules, and .agent-context/state/onboarding-report.json');
+    console.log(`- Repository workflows copied: no (workflows remain source-repo assets)`);
+    console.log(`- MCP template file: ${shouldIncludeMcpTemplate ? 'created (.vscode/mcp.json)' : 'not created by default (use --mcp-template)'}`);
     if (isTokenOptimizationEnabled) {
       console.log(`- Token optimization policy: enabled for ${selectedTokenAgentName}`);
     } else {
@@ -427,6 +438,7 @@ export async function runInitCommand(targetDirectoryArgument, initOptions = {})
     console.log('\nPlain-language summary:');
     console.log(`I prepared a ${selectedProfile.displayName.toLowerCase()} governance pack for a ${toTitleCase(selectedResolvedStackFileName)} project using the ${toTitleCase(selectedResolvedBlueprintFileName)} blueprint.`);
     console.log('Your AI tools will now receive one compiled rulebook plus the original source rules, and your review threshold is stored in .agent-context/policies/llm-judge-threshold.json.');
+    console.log('MCP server registration is manual inside your IDE settings, even when mcp.json exists.');
   } catch (error) {
     console.error('\n[FATAL] An error occurred during initialization. Attempting automatic rollback...');
     try {

package/lib/cli/commands/mcp.mjs

@@ -0,0 +1,3 @@
+export async function runMcpServerCommand() {
+  await import('../../../scripts/mcp-server.mjs');
+}

package/lib/cli/commands/upgrade.mjs

@@ -43,6 +43,7 @@ export function parseUpgradeArguments(commandArguments) {
     targetDirectory: '.',
     dryRun: false,
     skipConfirmation: false,
+    includeMcpTemplate: false,
   };
 
   for (let argumentIndex = 0; argumentIndex < commandArguments.length; argumentIndex++) {
@@ -63,6 +64,11 @@ export function parseUpgradeArguments(commandArguments) {
       continue;
     }
 
+    if (currentArgument === '--mcp-template') {
+      parsedUpgradeOptions.includeMcpTemplate = true;
+      continue;
+    }
+
     throw new Error(`Unknown option: ${currentArgument}`);
   }
 
@@ -95,7 +101,9 @@ export async function runUpgradeCommand(targetDirectoryArgument, upgradeOptions
     console.log(`\nAgentic-Senior-Core CLI v${CLI_VERSION}`);
     console.log('Running upgrade assistant for an existing repository.');
 
-    await copyGovernanceAssetsToTarget(resolvedTargetDirectoryPath);
+    await copyGovernanceAssetsToTarget(resolvedTargetDirectoryPath, {
+      includeMcpTemplate: upgradeOptions.includeMcpTemplate === true,
+    });
 
     const stackFileNames = await collectFileNames(path.join(AGENT_CONTEXT_DIR, 'stacks'));
     const blueprintFileNames = await collectFileNames(path.join(AGENT_CONTEXT_DIR, 'blueprints'));

package/lib/cli/constants.mjs

@@ -164,8 +164,8 @@ export const entryPointFiles = [
   '.cursorrules',
   '.windsurfrules',
   'AGENTS.md',
+  '.github/copilot-instructions.md',
   '.agent-override.md',
-  'mcp.json',
 ];
 
-export const directoryCopies = ['.agent-context', '.github', '.gemini', '.agents'];
+export const directoryCopies = ['.agent-context', '.gemini', '.agents'];

package/lib/cli/utils.mjs

@@ -29,8 +29,9 @@ export function printUsage() {
   console.log('Usage:');
   console.log('  agentic-senior-core launch');
   console.log('  agentic-senior-core init [target-directory] [--preset <name>] [--profile <beginner|balanced|strict>] [--profile-pack <name>] [--stack <name>] [--blueprint <name>] [--ci <true|false>] [--newbie] [--token-optimize] [--no-token-optimize] [--token-agent <name>]');
-  console.log('  agentic-senior-core upgrade [target-directory] [--dry-run] [--yes]');
+  console.log('  agentic-senior-core upgrade [target-directory] [--dry-run] [--yes] [--mcp-template]');
   console.log('  agentic-senior-core optimize [target-directory] [--agent <copilot|claude|cursor|windsurf|gemini|codex|cline>] [--enable|--disable] [--show]');
+  console.log('  agentic-senior-core mcp');
   console.log('  agentic-senior-core rollback [target-directory]');
   console.log('  agentic-senior-core skill [domain] [--tier <standard|advance|expert|above>] [--json]');
   console.log('  agentic-senior-core --version');
@@ -48,6 +49,7 @@ export function printUsage() {
   console.log('  --token-optimize  Explicitly enable token optimization policy during init (default behavior)');
   console.log('  --token-agent     Set token optimization agent target (copilot, claude, cursor, windsurf, gemini, codex, cline)');
   console.log('  --no-token-optimize  Disable token optimization policy during init');
+  console.log('  --mcp-template   Create .vscode/mcp.json workspace template (MCP trust/start remains manual in IDE)');
   console.log('  --dry-run    Preview upgrade without writing files');
   console.log('  --yes        Skip confirmation prompts for upgrade');
   console.log('  --agent      Target agent integration for token optimization mode');
@@ -96,7 +98,12 @@ export async function copyDirectory(sourceDirectoryPath, targetDirectoryPath) {
   }
 }
 
-export async function copyGovernanceAssetsToTarget(resolvedTargetDirectoryPath) {
+export async function copyGovernanceAssetsToTarget(
+  resolvedTargetDirectoryPath,
+  options = {}
+) {
+  const shouldIncludeMcpTemplate = options.includeMcpTemplate === true;
+
   for (const sourceDirectoryName of directoryCopies) {
     const sourceDirectoryPath = path.join(REPO_ROOT, sourceDirectoryName);
     if (!(await pathExists(sourceDirectoryPath))) {
@@ -121,6 +128,31 @@ export async function copyGovernanceAssetsToTarget(resolvedTargetDirectoryPath)
     await ensureDirectory(path.dirname(targetFilePath));
     await fs.copyFile(sourceFilePath, targetFilePath);
   }
+
+  if (shouldIncludeMcpTemplate) {
+    const vscodeDirectoryPath = path.join(resolvedTargetDirectoryPath, '.vscode');
+    const workspaceMcpConfigurationPath = path.join(vscodeDirectoryPath, 'mcp.json');
+
+    const workspaceMcpConfiguration = {
+      $schema: 'vscode://schemas/mcp',
+      servers: {
+        'agentic-senior-core': {
+          type: 'stdio',
+          command: 'npx',
+          args: ['-y', '@ryuenn3123/agentic-senior-core', 'mcp'],
+        },
+      },
+    };
+
+    if (!(await pathExists(workspaceMcpConfigurationPath))) {
+      await ensureDirectory(vscodeDirectoryPath);
+      await fs.writeFile(
+        workspaceMcpConfigurationPath,
+        JSON.stringify(workspaceMcpConfiguration, null, 2) + '\n',
+        'utf8'
+      );
+    }
+  }
 }
 
 export async function askChoice(promptMessage, options, userInterface) {

package/mcp.json

@@ -1,5 +1,4 @@
 {
-  "$schema": "https://modelcontextprotocol.io/schemas/mcp.json",
   "version": "1.0",
   "name": "agentic-senior-core",
   "description": "MCP configuration for governance-aware diagnostics and self-healing workflows with full knowledge injection.",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ryuenn3123/agentic-senior-core",
-  "version": "2.0.7",
+  "version": "2.0.9",
   "type": "module",
   "description": "Force your AI Agent to code like a Staff Engineer, not a Junior.",
   "bin": {

package/scripts/mcp-server.mjs

@@ -0,0 +1,347 @@
+#!/usr/bin/env node
+
+import { readFileSync } from 'node:fs';
+import { spawn } from 'node:child_process';
+import { dirname, resolve } from 'node:path';
+import { fileURLToPath } from 'node:url';
+
+const SCRIPT_FILE_PATH = fileURLToPath(import.meta.url);
+const REPOSITORY_ROOT = resolve(dirname(SCRIPT_FILE_PATH), '..');
+const PACKAGE_VERSION = JSON.parse(
+  readFileSync(resolve(REPOSITORY_ROOT, 'package.json'), 'utf8')
+).version;
+const DEFAULT_PROTOCOL_VERSION = '2024-11-05';
+
+const TEST_SUITE_ARGS = {
+  full: ['--test', './tests/cli-smoke.test.mjs', './tests/llm-judge.test.mjs', './tests/enterprise-ops.test.mjs', './tests/skill-tier-gate.test.mjs'],
+  cli: ['--test', './tests/cli-smoke.test.mjs'],
+  enterprise: ['--test', './tests/enterprise-ops.test.mjs'],
+  'llm-judge': ['--test', './tests/llm-judge.test.mjs'],
+  'skill-tier': ['--test', './tests/skill-tier-gate.test.mjs'],
+};
+
+const TOOL_DEFINITIONS = [
+  {
+    name: 'validate',
+    description: 'Run repository validation checks.',
+    inputSchema: {
+      type: 'object',
+      properties: {},
+      additionalProperties: false,
+    },
+  },
+  {
+    name: 'test',
+    description: 'Run test suites (full or targeted).',
+    inputSchema: {
+      type: 'object',
+      properties: {
+        suite: {
+          type: 'string',
+          enum: ['full', 'cli', 'enterprise', 'llm-judge', 'skill-tier'],
+          description: 'Target test suite. Defaults to full.',
+        },
+      },
+      additionalProperties: false,
+    },
+  },
+  {
+    name: 'release_gate',
+    description: 'Run release gate checks.',
+    inputSchema: {
+      type: 'object',
+      properties: {},
+      additionalProperties: false,
+    },
+  },
+  {
+    name: 'forbidden_content_check',
+    description: 'Run forbidden content scan used by publish gate.',
+    inputSchema: {
+      type: 'object',
+      properties: {},
+      additionalProperties: false,
+    },
+  },
+];
+
+let incomingBuffer = Buffer.alloc(0);
+
+function writeMessage(payload) {
+  const serializedPayload = JSON.stringify(payload);
+  const payloadLength = Buffer.byteLength(serializedPayload, 'utf8');
+  process.stdout.write(`Content-Length: ${payloadLength}\r\n\r\n${serializedPayload}`);
+}
+
+function sendResponse(id, result) {
+  writeMessage({
+    jsonrpc: '2.0',
+    id,
+    result,
+  });
+}
+
+function sendError(id, code, message, data) {
+  writeMessage({
+    jsonrpc: '2.0',
+    id,
+    error: {
+      code,
+      message,
+      data,
+    },
+  });
+}
+
+function normalizeToolName(rawToolName) {
+  return typeof rawToolName === 'string' ? rawToolName.trim() : '';
+}
+
+function buildCommandOutput(commandLabel, commandArguments, exitCode, stdoutContent, stderrContent) {
+  const outputSections = [
+    `Command: node ${commandArguments.join(' ')}`,
+    `Exit code: ${exitCode}`,
+  ];
+
+  if (stdoutContent.trim().length > 0) {
+    outputSections.push(`STDOUT:\n${stdoutContent.trimEnd()}`);
+  }
+
+  if (stderrContent.trim().length > 0) {
+    outputSections.push(`STDERR:\n${stderrContent.trimEnd()}`);
+  }
+
+  return [
+    `[${commandLabel}]`,
+    outputSections.join('\n\n'),
+  ].join('\n\n');
+}
+
+function runNodeCommand(commandLabel, commandArguments) {
+  return new Promise((resolveResult) => {
+    const childProcess = spawn(process.execPath, commandArguments, {
+      cwd: REPOSITORY_ROOT,
+      env: process.env,
+    });
+
+    let stdoutContent = '';
+    let stderrContent = '';
+
+    childProcess.stdout.on('data', (chunk) => {
+      stdoutContent += chunk.toString('utf8');
+    });
+
+    childProcess.stderr.on('data', (chunk) => {
+      stderrContent += chunk.toString('utf8');
+    });
+
+    childProcess.on('error', (error) => {
+      resolveResult({
+        content: [
+          {
+            type: 'text',
+            text: `[${commandLabel}] Failed to start command: ${error.message}`,
+          },
+        ],
+        isError: true,
+      });
+    });
+
+    childProcess.on('close', (exitCode) => {
+      const normalizedExitCode = typeof exitCode === 'number' ? exitCode : 1;
+      resolveResult({
+        content: [
+          {
+            type: 'text',
+            text: buildCommandOutput(
+              commandLabel,
+              commandArguments,
+              normalizedExitCode,
+              stdoutContent,
+              stderrContent
+            ),
+          },
+        ],
+        isError: normalizedExitCode !== 0,
+      });
+    });
+  });
+}
+
+async function executeToolCall(toolName, toolArguments = {}) {
+  if (toolName === 'validate') {
+    return runNodeCommand('validate', ['./scripts/validate.mjs']);
+  }
+
+  if (toolName === 'test') {
+    const requestedSuite = typeof toolArguments.suite === 'string'
+      ? toolArguments.suite
+      : 'full';
+
+    const selectedSuite = TEST_SUITE_ARGS[requestedSuite] ? requestedSuite : 'full';
+    return runNodeCommand(`test:${selectedSuite}`, TEST_SUITE_ARGS[selectedSuite]);
+  }
+
+  if (toolName === 'release_gate') {
+    return runNodeCommand('release_gate', ['./scripts/release-gate.mjs']);
+  }
+
+  if (toolName === 'forbidden_content_check') {
+    return runNodeCommand('forbidden_content_check', ['./scripts/forbidden-content-check.mjs']);
+  }
+
+  return {
+    content: [
+      {
+        type: 'text',
+        text: `Unknown tool: ${toolName}`,
+      },
+    ],
+    isError: true,
+  };
+}
+
+async function handleRequest(requestMessage) {
+  const requestId = requestMessage.id;
+  const requestMethod = requestMessage.method;
+  const requestParams = requestMessage.params || {};
+
+  if (typeof requestMethod !== 'string') {
+    if (typeof requestId !== 'undefined') {
+      sendError(requestId, -32600, 'Invalid Request');
+    }
+    return;
+  }
+
+  if (requestMethod === 'initialize') {
+    const negotiatedProtocolVersion = typeof requestParams.protocolVersion === 'string'
+      ? requestParams.protocolVersion
+      : DEFAULT_PROTOCOL_VERSION;
+
+    sendResponse(requestId, {
+      protocolVersion: negotiatedProtocolVersion,
+      capabilities: {
+        tools: {
+          listChanged: false,
+        },
+      },
+      serverInfo: {
+        name: 'agentic-senior-core',
+        version: PACKAGE_VERSION,
+      },
+    });
+    return;
+  }
+
+  if (requestMethod === 'notifications/initialized') {
+    return;
+  }
+
+  if (requestMethod === 'ping') {
+    if (typeof requestId !== 'undefined') {
+      sendResponse(requestId, {});
+    }
+    return;
+  }
+
+  if (requestMethod === 'tools/list') {
+    sendResponse(requestId, {
+      tools: TOOL_DEFINITIONS,
+    });
+    return;
+  }
+
+  if (requestMethod === 'tools/call') {
+    const requestedToolName = normalizeToolName(requestParams.name);
+
+    if (!requestedToolName) {
+      sendError(requestId, -32602, 'Invalid params: tool name is required');
+      return;
+    }
+
+    const toolResult = await executeToolCall(requestedToolName, requestParams.arguments || {});
+    sendResponse(requestId, toolResult);
+    return;
+  }
+
+  if (typeof requestId !== 'undefined') {
+    sendError(requestId, -32601, `Method not found: ${requestMethod}`);
+  }
+}
+
+function readNextFramedMessage() {
+  const headerEndIndex = incomingBuffer.indexOf('\r\n\r\n');
+  if (headerEndIndex === -1) {
+    return null;
+  }
+
+  const rawHeader = incomingBuffer.slice(0, headerEndIndex).toString('utf8');
+  const headerLines = rawHeader.split('\r\n');
+  let contentLength = null;
+
+  for (const headerLine of headerLines) {
+    const separatorIndex = headerLine.indexOf(':');
+    if (separatorIndex === -1) {
+      continue;
+    }
+
+    const headerName = headerLine.slice(0, separatorIndex).trim().toLowerCase();
+    const headerValue = headerLine.slice(separatorIndex + 1).trim();
+
+    if (headerName === 'content-length') {
+      contentLength = Number.parseInt(headerValue, 10);
+      break;
+    }
+  }
+
+  if (!Number.isFinite(contentLength) || contentLength < 0) {
+    incomingBuffer = Buffer.alloc(0);
+    return null;
+  }
+
+  const bodyStartIndex = headerEndIndex + 4;
+  const frameEndIndex = bodyStartIndex + contentLength;
+
+  if (incomingBuffer.length < frameEndIndex) {
+    return null;
+  }
+
+  const rawMessage = incomingBuffer.slice(bodyStartIndex, frameEndIndex).toString('utf8');
+  incomingBuffer = incomingBuffer.slice(frameEndIndex);
+  return rawMessage;
+}
+
+function processIncomingBuffer() {
+  while (true) {
+    const framedMessage = readNextFramedMessage();
+    if (framedMessage === null) {
+      return;
+    }
+
+    let parsedRequest;
+    try {
+      parsedRequest = JSON.parse(framedMessage);
+    } catch {
+      continue;
+    }
+
+    Promise.resolve(handleRequest(parsedRequest)).catch((error) => {
+      if (typeof parsedRequest?.id !== 'undefined') {
+        sendError(parsedRequest.id, -32603, 'Internal error', String(error?.message || error));
+      }
+    });
+  }
+}
+
+process.stdin.on('data', (chunk) => {
+  incomingBuffer = Buffer.concat([incomingBuffer, chunk]);
+  processIncomingBuffer();
+});
+
+process.stdin.on('end', () => {
+  process.exit(0);
+});
+
+process.on('SIGINT', () => {
+  process.exit(0);
+});

package/scripts/validate.mjs

@@ -39,6 +39,33 @@ const THIN_ADAPTER_PATHS = [
   '.github/copilot-instructions.md',
   '.gemini/instructions.md',
 ];
+const FORMAL_ARTIFACT_PATHS = [
+  '.instructions.md',
+  'README.md',
+  'CHANGELOG.md',
+  'docs/deep_analysis_and_roadmap_backlog.md',
+  '.agent-context/rules/api-docs.md',
+  '.agent-context/review-checklists/pr-checklist.md',
+  '.agent-context/prompts/review-code.md',
+  '.agent-context/skills/review-quality.md',
+  'AGENTS.md',
+  '.github/copilot-instructions.md',
+  '.gemini/instructions.md',
+];
+const REQUIRED_HUMAN_WRITING_SNIPPETS = [
+  {
+    path: '.agent-context/rules/api-docs.md',
+    snippets: ['## Human Writing Standard (Mandatory)', 'No emoji in formal artifacts.'],
+  },
+  {
+    path: '.agent-context/review-checklists/pr-checklist.md',
+    snippets: ['No emoji in formal documentation or review summaries', 'Documentation uses plain English and avoids AI cliches'],
+  },
+  {
+    path: 'docs/deep_analysis_and_roadmap_backlog.md',
+    snippets: ['## Part 6: Documentation and Explanation Standards (Mandatory)', 'No emoji in formal artifacts. This is mandatory.'],
+  },
+];
 
 const validationResult = {
   passed: 0,
@@ -124,6 +151,7 @@ async function validateRequiredFiles() {
     'scripts/benchmark-gate.mjs',
     'scripts/benchmark-intelligence.mjs',
     'scripts/governance-weekly-report.mjs',
+    'scripts/mcp-server.mjs',
     'scripts/frontend-usability-audit.mjs',
     'scripts/release-gate.mjs',
     'scripts/generate-sbom.mjs',
@@ -148,6 +176,7 @@ async function validateRequiredFiles() {
     '.agent-context/state/benchmark-watchlist.json',
     '.agent-context/state/skill-platform.json',
     '.agent-context/skills/index.json',
+    '.vscode/mcp.json',
     '.github/workflows/release-gate.yml',
     '.github/workflows/sbom-compliance.yml',
     '.github/workflows/benchmark-intelligence.yml',
@@ -673,6 +702,8 @@ async function validateMcpConfiguration() {
   const mcpConfiguration = JSON.parse(await readTextFile(join(ROOT_DIR, 'mcp.json')));
   const lintServerCommand = mcpConfiguration.servers?.lint?.command;
   const testServerCommand = mcpConfiguration.servers?.test?.command;
+  const workspaceMcpConfiguration = JSON.parse(await readTextFile(join(ROOT_DIR, '.vscode', 'mcp.json')));
+  const workspaceServerConfig = workspaceMcpConfiguration.servers?.['agentic-senior-core'];
 
   if (lintServerCommand === 'node') {
     pass('MCP lint server uses Node');
@@ -685,6 +716,64 @@ async function validateMcpConfiguration() {
   } else {
     fail('MCP test server must use Node');
   }
+
+  if (workspaceMcpConfiguration.$schema === 'vscode://schemas/mcp') {
+    pass('Workspace MCP config uses trusted VS Code schema');
+  } else {
+    fail('Workspace MCP config must use $schema: vscode://schemas/mcp');
+  }
+
+  if (workspaceServerConfig?.command === 'node') {
+    pass('Workspace MCP server command uses Node');
+  } else {
+    fail('Workspace MCP server command must use Node');
+  }
+
+  if (Array.isArray(workspaceServerConfig?.args) && workspaceServerConfig.args.includes('./scripts/mcp-server.mjs')) {
+    pass('Workspace MCP server points to scripts/mcp-server.mjs');
+  } else {
+    fail('Workspace MCP server must include ./scripts/mcp-server.mjs argument');
+  }
+}
+
+async function validateHumanWritingGovernance() {
+  console.log('\nChecking human writing governance...');
+
+  const disallowedEmojiPattern = /[\u2705\u274C\u26A0\u{1F4CC}\u{1F536}\u{1F4CE}\u{1F534}\u{1F7E0}\u{1F7E1}\u{1F7E2}]/u;
+
+  for (const formalArtifactPath of FORMAL_ARTIFACT_PATHS) {
+    const absoluteFormalArtifactPath = join(ROOT_DIR, formalArtifactPath);
+
+    if (!(await fileExists(absoluteFormalArtifactPath))) {
+      fail(`Missing formal artifact for writing governance: ${formalArtifactPath}`);
+      continue;
+    }
+
+    const formalArtifactContent = await readTextFile(absoluteFormalArtifactPath);
+
+    if (disallowedEmojiPattern.test(formalArtifactContent)) {
+      fail(`${formalArtifactPath} contains disallowed emoji symbols in formal text`);
+    } else {
+      pass(`${formalArtifactPath} has no disallowed emoji symbols`);
+    }
+  }
+
+  for (const snippetRule of REQUIRED_HUMAN_WRITING_SNIPPETS) {
+    const absoluteRulePath = join(ROOT_DIR, snippetRule.path);
+    if (!(await fileExists(absoluteRulePath))) {
+      fail(`Missing writing governance source: ${snippetRule.path}`);
+      continue;
+    }
+
+    const writingRuleContent = await readTextFile(absoluteRulePath);
+    for (const requiredSnippet of snippetRule.snippets) {
+      if (writingRuleContent.includes(requiredSnippet)) {
+        pass(`${snippetRule.path} includes writing governance snippet: ${requiredSnippet}`);
+      } else {
+        fail(`${snippetRule.path} is missing writing governance snippet: ${requiredSnippet}`);
+      }
+    }
+  }
 }
 
 async function validateInstructionAdapters() {
@@ -848,6 +937,7 @@ async function main() {
   await validateVersionConsistency();
   await validateDocumentationFlow();
   await validateMcpConfiguration();
+  await validateHumanWritingGovernance();
   await validateInstructionAdapters();
   await validateTrustTierSchema();
   await validateEvidenceBundles();