nodebench-mcp 2.22.0 → 2.26.0

package/dist/tools/toolRegistry.js

@@ -2070,6 +2070,18 @@ const REGISTRY_ENTRIES = [
         },
         phase: "research",
     },
+    {
+        name: "ingest_dive_screenshots",
+        category: "ui_ux_dive",
+        tags: ["ui", "screenshot", "ingest", "import", "bulk", "disk", "gallery", "dive", "png", "jpg"],
+        quickRef: {
+            nextAction: "Screenshots ingested into dive session. View them in the dashboard or use dive_screenshot to capture new ones.",
+            nextTools: ["dive_screenshot", "tag_ui_bug", "end_component_flow", "get_dive_tree"],
+            methodology: "agentic_vision",
+            tip: "Scans a directory for PNG/JPG files and bulk-imports them into a dive session's screenshot gallery. Use after external Playwright captures.",
+        },
+        phase: "test",
+    },
     // ═══════════════════════════════════════════
     // UI/UX DIVE V2 — Deep interaction testing,
     // screenshots, design audit, backend links,
@@ -2235,6 +2247,70 @@ const REGISTRY_ENTRIES = [
         },
         phase: "ship",
     },
+    {
+        name: "open_dive_dashboard",
+        category: "ui_ux_dive_v2",
+        tags: ["ui", "dashboard", "dive", "flywheel", "browser", "local", "report", "overview", "session", "open", "visualization"],
+        quickRef: {
+            nextAction: "Dashboard is open. Continue the dive — the dashboard auto-refreshes every 5s to show live progress.",
+            nextTools: ["start_ui_dive", "dive_auto_discover", "dive_code_locate", "dive_fix_verify"],
+            methodology: "agentic_vision",
+            tip: "Opens a local web dashboard (port 6274) showing the full flywheel cycle: routes, components, bugs, fixes, tests, reviews. Like Serena MCP's local page but for UI dives.",
+        },
+        phase: "utility",
+    },
+    // ═══════════════════════════════════════════
+    // SKILL SELF-UPDATE PROTOCOL — Track rule
+    // file provenance, staleness, and resync
+    // ═══════════════════════════════════════════
+    {
+        name: "register_skill",
+        category: "skill_update",
+        tags: ["skill", "rule", "register", "source", "hash", "frontmatter", "provenance", "memory", "agents-md", "cursor", "windsurf", "update"],
+        quickRef: {
+            nextAction: "Skill registered. Use check_skill_freshness periodically to detect when source files change.",
+            nextTools: ["check_skill_freshness", "list_skills"],
+            methodology: "self_reinforced_learning",
+            tip: "Register every .md rule file (e.g. .windsurf/rules/, AGENTS.md) with its source files, triggers, and update instructions. Enables automatic staleness detection.",
+        },
+        phase: "verify",
+    },
+    {
+        name: "check_skill_freshness",
+        category: "skill_update",
+        tags: ["skill", "freshness", "stale", "hash", "check", "drift", "source", "detect", "sync", "update", "rule"],
+        quickRef: {
+            nextAction: "If stale skills found, follow their update_instructions then call sync_skill to record the resync.",
+            nextTools: ["sync_skill", "list_skills", "register_skill"],
+            methodology: "self_reinforced_learning",
+            tip: "Run at session start or after big code changes. Compares SHA-256 hashes of source files to detect drift. Auto-updates skill status in DB.",
+        },
+        phase: "verify",
+    },
+    {
+        name: "sync_skill",
+        category: "skill_update",
+        tags: ["skill", "sync", "resync", "update", "hash", "refresh", "frontmatter", "rule", "source", "stale"],
+        quickRef: {
+            nextAction: "Skill synced. Verify the updated skill file is correct, then continue with your task.",
+            nextTools: ["check_skill_freshness", "list_skills"],
+            methodology: "self_reinforced_learning",
+            tip: "Call AFTER you have read the changed source files and updated the skill .md content. This tool records the sync and updates the hash.",
+        },
+        phase: "verify",
+    },
+    {
+        name: "list_skills",
+        category: "skill_update",
+        tags: ["skill", "list", "status", "overview", "rule", "memory", "history", "sync", "fresh", "stale"],
+        quickRef: {
+            nextAction: "Review skill statuses. Register any untracked rule files, check freshness for stale ones.",
+            nextTools: ["register_skill", "check_skill_freshness", "sync_skill"],
+            methodology: "self_reinforced_learning",
+            tip: "Use includeHistory:true to see the full sync timeline for each skill. Filter by status:'stale' to focus on what needs updating.",
+        },
+        phase: "utility",
+    },
     // ═══════════════════════════════════════════
     // MCP BRIDGE — Connect external MCP servers
     // ═══════════════════════════════════════════
@@ -2333,6 +2409,57 @@ const REGISTRY_ENTRIES = [
         },
         phase: "implement",
     },
+    // ═══════════════════════════════════════════
+    // QA ORCHESTRATION — Overstory multi-agent QA
+    // ═══════════════════════════════════════════
+    {
+        name: "overstory_fleet_status",
+        category: "qa_orchestration",
+        tags: ["overstory", "agent", "fleet", "status", "health", "multi-agent", "orchestration", "qa", "dogfood", "worktree"],
+        quickRef: {
+            nextAction: "Review agent states. If agents are idle, run dogfood:overstory to start a QA session.",
+            nextTools: ["overstory_qa_summary", "overstory_mail_log", "run_visual_qa_suite"],
+            methodology: "ai_flywheel",
+            tip: "Reads .overstory/agent-manifest.json and overstory.db. Shows configured agents, capabilities, gate policy, and live agent health.",
+        },
+        phase: "utility",
+    },
+    {
+        name: "overstory_qa_summary",
+        category: "qa_orchestration",
+        tags: ["overstory", "qa", "gate", "summary", "stability", "grade", "ssim", "triage", "p0", "p1", "dogfood"],
+        quickRef: {
+            nextAction: "If gate fails, check failing routes and fix p0/p1 issues. If gate passes, proceed to merge.",
+            nextTools: ["overstory_mail_log", "overstory_fleet_status", "run_visual_qa_suite", "burst_capture"],
+            methodology: "ai_flywheel",
+            tip: "Aggregates SSIM stability grades from visual_qa_runs and Gemini QA triage from Overstory mail. Returns gate pass/fail verdict.",
+        },
+        phase: "verify",
+    },
+    {
+        name: "overstory_mail_log",
+        category: "qa_orchestration",
+        tags: ["overstory", "mail", "log", "message", "route", "triage", "dispatch", "agent", "coordination"],
+        quickRef: {
+            nextAction: "Review messages to understand QA session state. Filter by type or agent for focused view.",
+            nextTools: ["overstory_qa_summary", "overstory_fleet_status", "overstory_merge_queue"],
+            methodology: "ai_flywheel",
+            tip: "Supports type_filter (result/dispatch/worker_done/escalation) and agent_filter. Shows structured mail payloads from the QA agent fleet.",
+        },
+        phase: "utility",
+    },
+    {
+        name: "overstory_merge_queue",
+        category: "qa_orchestration",
+        tags: ["overstory", "merge", "queue", "branch", "conflict", "gate", "builder", "qa", "resolution"],
+        quickRef: {
+            nextAction: "If branches are blocked, check QA gate failures. If pending, trigger merge with overstory merge --all.",
+            nextTools: ["overstory_qa_summary", "overstory_mail_log", "overstory_fleet_status"],
+            methodology: "ai_flywheel",
+            tip: "Shows FIFO merge queue with conflict resolution tiers. Use include_completed:true to see merge history.",
+        },
+        phase: "utility",
+    },
 ];
 // ── Exported lookup structures ───────────────────────────────────────────
 /** Map of tool name → registry entry for O(1) lookup */
@@ -2390,6 +2517,7 @@ const CATEGORY_COMPLEXITY = {
     email: "medium",
     rss: "low",
     architect: "low",
+    qa_orchestration: "low",
 };
 /** Per-tool complexity overrides (when category default is wrong) */
 const TOOL_COMPLEXITY_OVERRIDES = {
@@ -3190,6 +3318,7 @@ export const WORKFLOW_CHAINS = {
             { tool: "run_mandatory_flywheel", action: "6-step final verification" },
             { tool: "record_learning", action: "Capture what you learned" },
             { tool: "promote_to_eval", action: "Feed into eval batch" },
+            { tool: "save_session_note", action: "Save traceability note — cite original request, summarize what was delivered" },
         ],
     },
     fix_bug: {
@@ -3202,6 +3331,7 @@ export const WORKFLOW_CHAINS = {
             { tool: "log_test_result", action: "Record regression test" },
             { tool: "run_mandatory_flywheel", action: "6-step verification" },
             { tool: "record_learning", action: "Record the gotcha/pattern" },
+            { tool: "save_session_note", action: "Save traceability note — cite original request, record root cause and fix" },
         ],
     },
     ui_change: {
@@ -3215,6 +3345,7 @@ export const WORKFLOW_CHAINS = {
             { tool: "run_quality_gate", action: "Run ui_ux_qa gate" },
             { tool: "run_mandatory_flywheel", action: "Final verification" },
             { tool: "record_learning", action: "Record UI patterns" },
+            { tool: "save_session_note", action: "Save traceability note — cite original request, record visual evidence path" },
         ],
     },
     parallel_project: {
@@ -3532,5 +3663,52 @@ export const WORKFLOW_CHAINS = {
             { tool: "save_session_note", action: "Log sent emails so you have an audit trail that survives compaction" },
         ],
     },
+    webmcp_discovery: {
+        name: "WebMCP Origin Discovery",
+        description: "Connect to a WebMCP-enabled origin, discover its tools, and invoke them from the agent",
+        steps: [
+            { tool: "connect_webmcp_origin", action: "Connect to the target origin URL and establish a WebMCP session" },
+            { tool: "list_webmcp_tools", action: "List all tools exposed by the origin with schemas and annotations" },
+            { tool: "call_webmcp_tool", action: "Invoke a specific tool on the remote origin with arguments" },
+            { tool: "disconnect_webmcp_origin", action: "Clean up the WebMCP session when done" },
+        ],
+    },
+    batch_autopilot: {
+        name: "Batch Autopilot Run",
+        description: "Set up an operator profile and run a batch autopilot session for autonomous agent tasks",
+        steps: [
+            { tool: "setup_operator_profile", action: "Create or update USER.md and operator profile for autopilot context" },
+            { tool: "get_autopilot_status", action: "Check current autopilot readiness, profile completeness, and last run status" },
+            { tool: "trigger_batch_run", action: "Start a batch autopilot run using the operator profile as context" },
+            { tool: "get_batch_run_history", action: "Review history of past batch runs, outcomes, and timing" },
+            { tool: "sync_operator_profile", action: "Sync operator profile state from disk after manual edits" },
+        ],
+    },
+    daily_review: {
+        name: "Daily Brief Review",
+        description: "Pull the latest daily brief, review narrative threads, check ops dashboard, and sync to local storage",
+        steps: [
+            { tool: "sync_daily_brief", action: "Pull today's brief and narrative from Convex into local SQLite" },
+            { tool: "get_daily_brief_summary", action: "Get the full brief summary with key signals and insights" },
+            { tool: "get_narrative_status", action: "Check narrative thread status — dominant story, under-reported angle, evidence scores" },
+            { tool: "get_ops_dashboard", action: "Review pipeline health: posting status, tool usage, active workflows" },
+            { tool: "open_local_dashboard", action: "Open the local HTML dashboard in the browser for visual review" },
+        ],
+    },
+    gemini_qa: {
+        name: "Gemini Vision QA Loop",
+        description: "Automated UI/UX quality gate — capture screenshots (dark/light × desktop/mobile), send to Gemini Flash for Jony Ive product design review, fix issues, loop until 100/100",
+        steps: [
+            { tool: "check_mcp_setup", action: "Verify Gemini API key (GOOGLE_AI_KEY) and vision domain are ready" },
+            { tool: "start_verification_cycle", action: "Open a verification cycle titled 'Gemini QA Loop' to track progress" },
+            { tool: "save_session_note", action: "Shell: `npx vite build` then `npx playwright test tests/e2e/full-ui-dogfood.spec.ts --project=chromium --workers=1` — capture 4-variant screenshots" },
+            { tool: "save_session_note", action: "Shell: `npm run dogfood:publish` — copy screenshots to public/dogfood/ with variant metadata manifest" },
+            { tool: "save_session_note", action: "Shell: `npx vite build && node scripts/ui/runDogfoodGeminiQa.mjs` — rebuild, launch preview, trigger Gemini QA" },
+            { tool: "log_test_result", action: "Log QA score from public/dogfood/qa-results.json — formula: 100 - P1×6 - P2×2 - P3×1" },
+            { tool: "save_session_note", action: "Fix P1 issues (6pts each) then P2 (2pts) then P3 (1pt) — root-cause each before fixing" },
+            { tool: "get_overstory_qa_gate", action: "Check QA gate for per-route stability grades and issue counts" },
+            { tool: "record_learning", action: "Record QA trajectory and Gemini finding patterns for regression tracking" },
+        ],
+    },
 };
 //# sourceMappingURL=toolRegistry.js.map