delimit-cli 3.14.14 → 3.14.16

package/adapters/cursor-rules.js

@@ -11,6 +11,9 @@
 const fs = require('fs');
 const path = require('path');
 
+// LED-213: Import canonical template for cross-model parity
+const { getDelimitSection } = require('../lib/delimit-template');
+
 const HOME = process.env.HOME || '';
 const CURSOR_DIR = path.join(HOME, '.cursor');
 const CURSOR_RULES_DIR = path.join(CURSOR_DIR, 'rules');
@@ -50,35 +53,8 @@ function uninstallRules() {
 }
 
 function getDelimitRules(version) {
-    return `# Delimit Governance (v${version})
-
-## On every session start:
-1. Call \`delimit_ledger_context\` to check for open tasks
-2. Call \`delimit_gov_health\` to check governance status
-
-## After editing code:
-- After editing UI/CSS: call \`delimit_design_validate_responsive\`
-- After editing API specs: call \`delimit_lint\`
-- After editing tests: call \`delimit_test_smoke\`
-
-## Before deploying:
-1. Call \`delimit_security_audit\`
-2. Call \`delimit_test_smoke\`
-3. Call \`delimit_deploy_plan\`
-
-## Before committing:
-- Call \`delimit_repo_diagnose\` to check for common issues
-
-## Governance enforcement:
-- All tool calls route through governance automatically (server-side)
-- Critical actions are blocked until approved
-- High-risk actions require approval in enforce mode
-- Check mode: advisory (warn only), guarded (block critical), enforce (block critical + high-risk)
-
-## Links
-- Docs: https://delimit.ai/docs
-- GitHub: https://github.com/delimit-ai/delimit-mcp-server
-`;
+    // LED-213: Use canonical Consensus 123 template for Cursor parity
+    return getDelimitSection();
 }
 
 module.exports = { installRules, uninstallRules, getDelimitRules };

package/bin/delimit-setup.js

@@ -839,8 +839,62 @@ exit 127
     }
     log('');
 
-    // Step 11: Done
-    step(11, 'Done!');
+    // Step 11: Social target scanning config
+    step(11, 'Configuring social target scanner...');
+
+    const socialConfigPath = path.join(DELIMIT_HOME, 'social_target_config.json');
+    const socialDefaultConfig = {
+        platforms: {
+            x: { enabled: true, provider: 'twttr241' },
+            reddit: { enabled: true, provider: 'proxy' },
+            github: { enabled: true, provider: 'gh_cli' },
+            hn: { enabled: true, provider: 'algolia' },
+            devto: { enabled: true, provider: 'public_api' },
+            namepros: { enabled: false, provider: 'manual' },
+        },
+        subreddits: {},
+        github_queries: {},
+        scan_limit: 10,
+        min_engagement: { score: 1, comments: 2 },
+    };
+
+    if (!fs.existsSync(socialConfigPath)) {
+        fs.mkdirSync(path.dirname(socialConfigPath), { recursive: true });
+        fs.writeFileSync(socialConfigPath, JSON.stringify(socialDefaultConfig, null, 2) + '\n');
+        log(`  ${green('\u2713')} Created ${dim(socialConfigPath)}`);
+    } else {
+        log(`  ${dim('  Config already exists:')} ${socialConfigPath}`);
+    }
+
+    // Auto-detect available platforms
+    const detectedPlatforms = {};
+    // HN and Dev.to are always available (public APIs)
+    detectedPlatforms['hn'] = 'available (public API)';
+    detectedPlatforms['devto'] = 'available (public API)';
+    // GitHub: check gh CLI
+    try {
+        execSync('gh auth status', { stdio: 'pipe', timeout: 10000 });
+        detectedPlatforms['github'] = 'available (gh authenticated)';
+    } catch {
+        detectedPlatforms['github'] = 'unavailable (gh not authenticated)';
+    }
+    // X: check for RapidAPI key in env
+    if (process.env.RAPIDAPI_KEY) {
+        detectedPlatforms['x'] = 'available (RapidAPI key in env)';
+    } else if (process.env.XAI_API_KEY) {
+        detectedPlatforms['x'] = 'available (xAI key in env, fallback)';
+    } else {
+        detectedPlatforms['x'] = 'unavailable (no API key)';
+    }
+
+    for (const [plat, status] of Object.entries(detectedPlatforms)) {
+        const icon = status.startsWith('available') ? green('\u2713') : yellow('\u2717');
+        log(`  ${icon} ${plat}: ${dim(status)}`);
+    }
+    log('');
+
+    // Step 12: Done
+    step(12, 'Done!');
     log('');
     log(`  ${green('Delimit is installed.')} Your AI now has persistent memory and governance.`);
     log('');
@@ -872,87 +926,8 @@ exit 127
     log('');
 }
 
-function getDelimitSection() {
-    const pkg = JSON.parse(fs.readFileSync(path.join(__dirname, '..', 'package.json'), 'utf-8'));
-    const version = pkg.version || '0.0.0';
-    return `<!-- delimit:start v${version} -->
-# Delimit
-
-Unify all AI coding assistants with persistent context, governance, and multi-model debate.
-
-## Auto-Trigger Rules (Consensus 123)
-
-These rules fire automatically. The AI MUST call the specified tools without asking permission.
-The founder reviews and approves via email. The AI's job is to draft, send, and chain — not to ask.
-
-### Session Lifecycle
-- **Session start**: call \`delimit_ledger_context\` + \`delimit_gov_health\` + \`delimit_inbox_daemon(action="start")\` + \`delimit_notify_routing(action="status")\`
-- **Session end or idle**: call \`delimit_session_handoff\` to preserve context for next session
-- If returning user, summarize what's pending. If new project, call \`delimit_scan\`.
-
-### Code Development (debounced per edit batch)
-- After editing UI/CSS: call \`delimit_design_validate_responsive\`
-- After editing API specs: call \`delimit_lint\` + \`delimit_drift_check\`
-- After editing tests: call \`delimit_test_smoke\`
-- After any significant code change: call \`delimit_test_smoke\` if applicable tests exist
-
-### Pre-Commit / Pre-Deploy Gates
-- Before \`git commit\`: call \`delimit_repo_diagnose\`
-- Deploy requested: \`delimit_security_audit\` then \`delimit_test_smoke\` then \`delimit_changelog\` then \`delimit_deploy_plan\`
-- Deploy succeeds: \`delimit_deploy_verify\` + \`delimit_evidence_collect\`
-- Verify succeeds: \`delimit_ledger_done\` (related task) + \`delimit_notify\`
-
-### Content / Outreach (all drafts, never auto-post)
-- Social content drafted: ALWAYS call \`delimit_social_post\` or \`delimit_notify\` immediately
-- \`delimit_social_target\` scan completes: call \`delimit_notify\` with target summary
-- Screenshot of social thread shared: auto-draft reply + call \`delimit_notify\`
-
-### Strategic / Business Operations
-- Strategic opportunity identified: call \`delimit_ledger_add\` immediately
-- \`delimit_ledger_add\` fires (non-P0): \`delimit_notify\` (batched, 15-min digest)
-- \`delimit_ledger_add\` fires (P0): \`delimit_notify\` (immediate) + \`delimit_agent_dispatch\`
-- Conflicting high-severity signals: call \`delimit_deliberate\` for multi-model consensus
-
-### Monitoring / Sensors
-- GitHub issue state change (closed, merged, labeled): \`delimit_ledger_update\` + \`delimit_notify\` (immediate)
-- GitHub issue new comments: \`delimit_ledger_update\` + \`delimit_notify\` (batched digest)
-- \`delimit_drift_check\` finds drift: \`delimit_ledger_add\` + \`delimit_notify\`
-
-### Approval Loop (Email Control Plane)
-- Approval email detected: execute the approved action based on context
-- Rejection email detected: \`delimit_ledger_update\` (mark rejected)
-
-### Audit Trail
-- After security audit, test run, or deploy: call \`delimit_evidence_collect\`
-- Any gate failure: \`delimit_evidence_collect\` + \`delimit_ledger_add\` + \`delimit_notify\` (immediate)
-
-### Storm Prevention
-- Edit triggers: debounce to once per assistant turn
-- Non-P0 notifications: batch into 15-minute digests
-- Max 5 immediate notifications per hour (overflow to digest)
-- Per-scan fan-out: max 10 targets per notification
-
-## What Delimit does:
-- **API governance** -- lint, diff, semver classification, migration guides
-- **Persistent context** -- memory and ledger survive across sessions and models
-- **Security audit** -- dependency scanning, secret detection
-- **Test verification** -- confirm tests actually ran, measure coverage
-- **Multi-model deliberation** -- consensus from multiple AI models (3 free, then BYOK)
-
-## GitHub Action
-Add breaking change detection to any repo:
-\`\`\`yaml
-- uses: delimit-ai/delimit-action@v1
-  with:
-    spec: api/openapi.yaml
-\`\`\`
-
-## Links
-- Docs: https://delimit.ai/docs
-- GitHub: https://github.com/delimit-ai/delimit-mcp-server
-- Action: https://github.com/marketplace/actions/delimit-api-governance
-<!-- delimit:end -->`;
-}
+// LED-213: Import canonical template from shared module
+const { getDelimitSection } = require('../lib/delimit-template');
 
 function getClaudeMdContent() {
     return getDelimitSection() + '\n';

package/lib/cross-model-hooks.js

@@ -17,6 +17,9 @@ const path = require('path');
 const { execSync } = require('child_process');
 const os = require('os');
 
+// LED-213: Import canonical template for cross-model parity
+const { getDelimitSection, getDelimitSectionCondensed } = require('./delimit-template');
+
 // Use process.env.HOME to allow test overrides; fall back to os.homedir()
 function getHome() { return process.env.HOME || os.homedir(); }
 function getDelimitHome() { return path.join(getHome(), '.delimit'); }
@@ -348,25 +351,9 @@ function installCodexHooks(tool, hookConfig) {
     // Codex instructions.md -- acts as the session-start equivalent
     if (hookConfig.session_start) {
         const instructionsPath = tool.instructionsPath || path.join(codexDir, 'instructions.md');
+        // LED-213: Use canonical Consensus 123 template for Codex parity
         const delimitBlock = `<!-- delimit:hooks-start -->
-## Delimit Governance
-
-On every session start:
-1. Run this command to check governance status:
-\`\`\`
-npx delimit-cli hook session-start
-\`\`\`
-2. Call \`delimit_inbox_daemon(action="start")\` to start email polling for approvals.
-
-Before committing changes, run:
-\`\`\`
-npx delimit-cli hook pre-commit
-\`\`\`
-
-After editing OpenAPI specs, run:
-\`\`\`
-npx delimit-cli hook pre-tool Edit
-\`\`\`
+${getDelimitSection()}
 <!-- delimit:hooks-end -->`;
 
         let content = '';
@@ -427,39 +414,20 @@ function installGeminiHooks(tool, hookConfig) {
         } catch { config = {}; }
     }
 
-    const govInstructions = [
-        'Delimit governance is active.',
-        'On session start: (1) run: npx delimit-cli hook session-start,',
-        '(2) call delimit_inbox_daemon(action="start") to start email polling.',
-        'Before committing, run: npx delimit-cli hook pre-commit',
-        'After editing API specs, run: npx delimit-cli hook pre-tool Edit',
-        'Call delimit_ledger_context and delimit_gov_health via MCP when available.',
-    ].join(' ');
+    // LED-213: Use canonical Consensus 123 template (condensed for JSON)
+    const govInstructions = getDelimitSectionCondensed();
 
-    if (!config.customInstructions || !config.customInstructions.includes('delimit-cli hook')) {
+    if (!config.customInstructions || !config.customInstructions.includes('Consensus 123')) {
         config.customInstructions = govInstructions;
         changes.push('customInstructions');
     }
 
     fs.writeFileSync(tool.configPath, JSON.stringify(config, null, 2));
 
-    // Write a GEMINI.md governance file (equivalent of CLAUDE.md)
+    // LED-213: Write GEMINI.md with canonical Consensus 123 template
     const geminiMd = path.join(geminiDir, 'GEMINI.md');
-    if (!fs.existsSync(geminiMd) || !fs.readFileSync(geminiMd, 'utf-8').includes('delimit')) {
-        const content = `# Delimit Governance
-
-On every session start:
-1. Run \`npx delimit-cli hook session-start\` to check open tasks and governance status
-2. Call \`delimit_inbox_daemon(action="start")\` to start email polling for approvals
-
-After editing code:
-- After editing API specs: run \`npx delimit-cli hook pre-tool Edit\`
-- After editing tests: run \`npx delimit-cli hook pre-tool Edit\`
-
-Before committing:
-- Run \`npx delimit-cli hook pre-commit\` to check for issues
-`;
-        fs.writeFileSync(geminiMd, content);
+    if (!fs.existsSync(geminiMd) || !fs.readFileSync(geminiMd, 'utf-8').includes('Consensus 123')) {
+        fs.writeFileSync(geminiMd, getDelimitSection() + '\n');
         changes.push('GEMINI.md');
     }
 

package/lib/delimit-template.js

@@ -0,0 +1,111 @@
+/**
+ * LED-213: Canonical Delimit instruction template.
+ *
+ * Single source of truth for the governance section injected into
+ * CLAUDE.md, Codex instructions.md, Gemini GEMINI.md / customInstructions,
+ * and Cursor .cursorrules.  All models get identical Consensus 123 content.
+ */
+
+const fs = require('fs');
+const path = require('path');
+
+/**
+ * Return the full Consensus 123 governance template (markdown).
+ * Used verbatim by Claude, Codex, and Cursor.
+ */
+function getDelimitSection() {
+    const pkg = JSON.parse(fs.readFileSync(path.join(__dirname, '..', 'package.json'), 'utf-8'));
+    const version = pkg.version || '0.0.0';
+    return `<!-- delimit:start v${version} -->
+# Delimit
+
+Unify all AI coding assistants with persistent context, governance, and multi-model debate.
+
+## Auto-Trigger Rules (Consensus 123)
+
+These rules fire automatically. The AI MUST call the specified tools without asking permission.
+The founder reviews and approves via email. The AI's job is to draft, send, and chain — not to ask.
+
+### Session Lifecycle
+- **Session start**: call \`delimit_ledger_context\` + \`delimit_gov_health\` + \`delimit_inbox_daemon(action="start")\` + \`delimit_notify_routing(action="status")\`
+- **Session end or idle**: call \`delimit_session_handoff\` to preserve context for next session
+- If returning user, summarize what's pending. If new project, call \`delimit_scan\`.
+
+### Code Development (debounced per edit batch)
+- After editing UI/CSS: call \`delimit_design_validate_responsive\`
+- After editing API specs: call \`delimit_lint\` + \`delimit_drift_check\`
+- After editing tests: call \`delimit_test_smoke\`
+- After any significant code change: call \`delimit_test_smoke\` if applicable tests exist
+
+### Pre-Commit / Pre-Deploy Gates
+- Before \`git commit\`: call \`delimit_repo_diagnose\`
+- Deploy requested: \`delimit_security_audit\` then \`delimit_test_smoke\` then \`delimit_changelog\` then \`delimit_deploy_plan\`
+- Deploy succeeds: \`delimit_deploy_verify\` + \`delimit_evidence_collect\`
+- Verify succeeds: \`delimit_ledger_done\` (related task) + \`delimit_notify\`
+
+### Content / Outreach (all drafts, never auto-post)
+- Social content drafted: ALWAYS call \`delimit_social_post\` or \`delimit_notify\` immediately
+- \`delimit_social_target\` scan completes: call \`delimit_notify\` with target summary
+- Screenshot of social thread shared: auto-draft reply + call \`delimit_notify\`
+- During ANY social interaction: scan for competitive signals, integration opportunities, and strategic items. Add to strategy ledger if significant.
+
+### Strategic / Business Operations
+- Strategic opportunity identified: call \`delimit_ledger_add\` immediately
+- \`delimit_ledger_add\` fires (non-P0): \`delimit_notify\` (batched, 15-min digest)
+- \`delimit_ledger_add\` fires (P0): \`delimit_notify\` (immediate) + \`delimit_agent_dispatch\`
+- Conflicting high-severity signals: call \`delimit_deliberate\` for multi-model consensus
+
+### Monitoring / Sensors
+- GitHub issue state change (closed, merged, labeled): \`delimit_ledger_update\` + \`delimit_notify\` (immediate)
+- GitHub issue new comments: \`delimit_ledger_update\` + \`delimit_notify\` (batched digest)
+- \`delimit_drift_check\` finds drift: \`delimit_ledger_add\` + \`delimit_notify\`
+
+### Approval Loop (Email Control Plane)
+- Approval email detected: execute the approved action based on context
+- Rejection email detected: \`delimit_ledger_update\` (mark rejected)
+
+### Audit Trail
+- After security audit, test run, or deploy: call \`delimit_evidence_collect\`
+- Any gate failure: \`delimit_evidence_collect\` + \`delimit_ledger_add\` + \`delimit_notify\` (immediate)
+
+### Storm Prevention
+- Edit triggers: debounce to once per assistant turn
+- Non-P0 notifications: batch into 15-minute digests
+- Max 5 immediate notifications per hour (overflow to digest)
+- Per-scan fan-out: max 10 targets per notification
+
+## What Delimit does:
+- **API governance** -- lint, diff, semver classification, migration guides
+- **Persistent context** -- memory and ledger survive across sessions and models
+- **Security audit** -- dependency scanning, secret detection
+- **Test verification** -- confirm tests actually ran, measure coverage
+- **Multi-model deliberation** -- consensus from multiple AI models (3 free, then BYOK)
+
+## GitHub Action
+Add breaking change detection to any repo:
+\`\`\`yaml
+- uses: delimit-ai/delimit-action@v1
+  with:
+    spec: api/openapi.yaml
+\`\`\`
+
+## Links
+- Docs: https://delimit.ai/docs
+- GitHub: https://github.com/delimit-ai/delimit-mcp-server
+- Action: https://github.com/marketplace/actions/delimit-api-governance
+<!-- delimit:end -->`;
+}
+
+/**
+ * Return a condensed single-line version for Gemini customInstructions (JSON string value).
+ * Newlines are literal \\n so it fits inside a JSON string.
+ */
+function getDelimitSectionCondensed() {
+    return getDelimitSection()
+        .replace(/<!-- delimit:start[^>]*-->\n?/, '')
+        .replace(/<!-- delimit:end -->\n?/, '')
+        .replace(/\n/g, '\\n')
+        .trim();
+}
+
+module.exports = { getDelimitSection, getDelimitSectionCondensed };

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "delimit-cli",
   "mcpName": "io.github.delimit-ai/delimit-mcp-server",
-  "version": "3.14.14",
+  "version": "3.14.16",
   "description": "Unify Claude Code, Codex, Cursor, and Gemini CLI with persistent context, governance, and multi-model debate.",
   "main": "index.js",
   "files": [