nodebench-mcp 2.62.0 → 2.63.0

package/dist/benchmarks/llmJudgeEval.js

@@ -275,11 +275,11 @@ function founderTemplates() {
     return [
         // weekly_reset
         { query: "What changed in our product direction this week?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "Give me a weekly reset briefing for the founding team", scenario: "weekly_reset", expectedTools: ["founder_local_weekly_reset", "founder_deep_context_gather"], forbiddenTools: ["check_contract_compliance"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "Summarize last week's key decisions and their rationale", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["generate_zero_draft"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "What are the top 3 risks to our current sprint?", scenario: "weekly_reset", expectedTools: ["founder_deep_context_gather", "get_proactive_alerts"], forbiddenTools: ["run_recon", "check_page_performance"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output includes quantitative data points or metrics", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "How is our burn rate tracking against the runway?", scenario: "weekly_reset", expectedTools: ["founder_deep_context_gather"], forbiddenTools: ["run_recon", "check_email_setup"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output includes quantitative data points or metrics", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "What did we ship this week and what slipped?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["generate_report"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Give me a weekly reset briefing for the founding team", scenario: "weekly_reset", expectedTools: ["founder_local_weekly_reset", "founder_deep_context_gather"], forbiddenTools: ["founder_packet_validate"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Summarize last week's key decisions and their rationale", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["render_decision_memo"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "What are the top 3 risks to our current sprint?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output includes quantitative data points or metrics", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "How is our burn rate tracking against the runway?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon", "check_mcp_setup"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output includes quantitative data points or metrics", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "What did we ship this week and what slipped?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["export_artifact_packet"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         // company_search
         { query: "Research Stripe and tell me about their latest product moves", scenario: "company_search", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }, { criterion: "Output includes quantitative data points or metrics", weight: 1 }] },
         { query: "Pull everything you know about Anthropic's recent funding", scenario: "company_search", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
@@ -288,16 +288,16 @@ function founderTemplates() {
         { query: "What are the moats of our top 3 competitors?", scenario: "competitor_brief", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["check_mcp_setup"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         // delegation — route to founder_local_synthesize with pre_delegation packetType
         { query: "Draft a delegation brief for the engineering lead on the auth refactor", scenario: "delegation", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Tool returned valid structured JSON or object data", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "Create a handoff packet for the new VP of Product", scenario: "delegation", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["check_contract_compliance"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Tool returned valid structured JSON or object data", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Create a handoff packet for the new VP of Product", scenario: "delegation", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_packet_validate"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Tool returned valid structured JSON or object data", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         // important_change
-        { query: "Flag any important changes in our competitive landscape this week", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["start_verification_cycle"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Flag any important changes in our competitive landscape this week", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["start_dogfood_session"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "What's the most critical thing I should know about right now?", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         // memo_export — route to founder_local_weekly_reset which produces a full memo
-        { query: "Export our latest decision memo as a shareable packet", scenario: "memo_export", expectedTools: ["founder_local_weekly_reset"], forbiddenTools: ["run_recon", "check_page_performance"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Tool returned valid structured JSON or object data", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Export our latest decision memo as a shareable packet", scenario: "memo_export", expectedTools: ["founder_local_weekly_reset"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Tool returned valid structured JSON or object data", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "Package the Q1 strategy review for the board", scenario: "memo_export", expectedTools: ["founder_local_weekly_reset"], forbiddenTools: ["start_dogfood_session"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Tool returned valid structured JSON or object data", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         // packet_diff — route to founder_local_synthesize with important_change (shows what changed)
         { query: "What changed between the last two strategy packets?", scenario: "packet_diff", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Tool returned valid structured JSON or object data", weight: 1 }] },
-        { query: "Show me the delta between our January and March founder packets", scenario: "packet_diff", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["check_email_setup"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Tool returned valid structured JSON or object data", weight: 1 }] },
+        { query: "Show me the delta between our January and March founder packets", scenario: "packet_diff", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["check_mcp_setup"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Tool returned valid structured JSON or object data", weight: 1 }] },
         // role_switch
         { query: "Switch to investor mode and evaluate our pitch deck", scenario: "role_switch", expectedTools: ["founder_local_synthesize"], forbiddenTools: [], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }] },
         { query: "I need to think like a banker — what's the credit risk here?", scenario: "role_switch", expectedTools: ["founder_local_synthesize"], forbiddenTools: [], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }] },
@@ -312,7 +312,7 @@ function bankerTemplates() {
         { query: "Compare the credit profiles of Company A vs Company B", scenario: "competitor_brief", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "Draft a term sheet summary for the lending committee", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output structure matches the tool's documented schema", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "What's changed in the regulatory landscape this week?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "Export the due diligence findings for the Acme Corp loan", scenario: "memo_export", expectedTools: ["founder_local_synthesize", "get_recon_summary"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output structure matches the tool's documented schema", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Export the due diligence findings for the Acme Corp loan", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output structure matches the tool's documented schema", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "Show me how the risk ratings shifted since last quarter", scenario: "packet_diff", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Multiple tools in the chain produced non-empty results", weight: 1 }] },
         { query: "Delegate the annual review prep to the junior analyst", scenario: "delegation", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output structure matches the tool's documented schema", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "Assess the market risk exposure in our current book", scenario: "company_search", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }, { criterion: "Output includes quantitative data points or metrics", weight: 1 }] },
@@ -323,16 +323,16 @@ function bankerTemplates() {
 }
 function ceoTemplates() {
     return [
-        { query: "Give me the executive summary of where we stand this week", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["check_page_performance"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Give me the executive summary of where we stand this week", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: [], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "What should I be worried about that nobody's telling me?", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "Prepare talking points for the all-hands meeting", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["check_contract_compliance"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output structure matches the tool's documented schema", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Prepare talking points for the all-hands meeting", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_packet_validate"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output structure matches the tool's documented schema", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "How are our OKRs tracking this quarter?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output includes quantitative data points or metrics", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "Who on the leadership team needs my attention this week?", scenario: "delegation", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["check_mcp_setup"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output structure matches the tool's documented schema", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "Draft a board update email for this month", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output structure matches the tool's documented schema", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "What's our competitive position changed to since last month?", scenario: "competitor_brief", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "Compare the last two quarterly reviews for drift", scenario: "packet_diff", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["check_email_setup"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Multiple tools in the chain produced non-empty results", weight: 1 }] },
+        { query: "Compare the last two quarterly reviews for drift", scenario: "packet_diff", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["check_mcp_setup"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Multiple tools in the chain produced non-empty results", weight: 1 }] },
         { query: "I need to delegate the hiring pipeline review — create a brief", scenario: "delegation", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output structure matches the tool's documented schema", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "Switch to founder mode and deep-dive into the product roadmap", scenario: "role_switch", expectedTools: ["discover_tools"], forbiddenTools: ["check_contract_compliance"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }] },
+        { query: "Switch to founder mode and deep-dive into the product roadmap", scenario: "role_switch", expectedTools: ["discover_tools"], forbiddenTools: ["founder_packet_validate"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }] },
         { query: "Flag the most important thing that changed since yesterday", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "Research what our key enterprise customers are saying publicly", scenario: "company_search", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }, { criterion: "Output includes quantitative data points or metrics", weight: 1 }] },
     ];
@@ -340,24 +340,24 @@ function ceoTemplates() {
 function researcherTemplates() {
     return [
         { query: "Find recent papers on transformer attention mechanisms", scenario: "company_search", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }, { criterion: "Output includes quantitative data points or metrics", weight: 1 }] },
-        { query: "Build a research digest on federated learning advances in 2025", scenario: "company_search", expectedTools: ["build_research_digest", "run_recon"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Build a research digest on federated learning advances in 2025", scenario: "company_search", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "What are the open problems in RLHF that nobody's solved?", scenario: "competitor_brief", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_synthesize"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "Summarize the key findings from this week's arXiv papers on LLM reasoning", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "Compare the methodology of these two papers on knowledge distillation", scenario: "competitor_brief", expectedTools: ["compare_options", "build_research_digest"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Compare the methodology of these two papers on knowledge distillation", scenario: "competitor_brief", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "Export my literature review notes as a shareable document", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output structure matches the tool's documented schema", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "What contradictions exist in the current MoE literature?", scenario: "important_change", expectedTools: ["build_research_digest", "get_important_changes"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "Track how the consensus on scaling laws has shifted this year", scenario: "packet_diff", expectedTools: ["founder_local_synthesize", "build_research_digest"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Multiple tools in the chain produced non-empty results", weight: 1 }] },
+        { query: "What contradictions exist in the current MoE literature?", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Track how the consensus on scaling laws has shifted this year", scenario: "packet_diff", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Multiple tools in the chain produced non-empty results", weight: 1 }] },
         { query: "Delegate the data collection task to the research assistant", scenario: "delegation", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output structure matches the tool's documented schema", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "Switch to operator mode and check if the experiment pipeline is healthy", scenario: "role_switch", expectedTools: ["discover_tools"], forbiddenTools: ["build_research_digest"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }] },
+        { query: "Switch to operator mode and check if the experiment pipeline is healthy", scenario: "role_switch", expectedTools: ["discover_tools"], forbiddenTools: ["search_all_knowledge"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }] },
         { query: "What are the most-cited papers in agentic AI from 2025?", scenario: "company_search", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }, { criterion: "Output includes quantitative data points or metrics", weight: 1 }] },
-        { query: "Generate a research question from the gaps in current RAG literature", scenario: "important_change", expectedTools: ["build_research_digest"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Generate a research question from the gaps in current RAG literature", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
     ];
 }
 function studentTemplates() {
     return [
         { query: "Help me understand how transformers work at a high level", scenario: "company_search", expectedTools: ["discover_tools"], forbiddenTools: ["founder_deep_context_gather", "founder_local_synthesize"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "What should I study this week for my ML course?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon", "founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
-        { query: "Compare supervised vs unsupervised learning for my report", scenario: "competitor_brief", expectedTools: ["compare_options", "discover_tools"], forbiddenTools: ["run_recon", "founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
+        { query: "Compare supervised vs unsupervised learning for my report", scenario: "competitor_brief", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon", "founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "Export my study notes as a markdown document", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon", "founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output structure matches the tool's documented schema", weight: 1 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "What changed in the AI landscape this week that I should know about?", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains temporal information (dates, timestamps, periods)", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }, { criterion: "Output does not contain error stack traces or crash messages", weight: 1 }] },
         { query: "I need to switch to a research perspective for my thesis", scenario: "role_switch", expectedTools: ["founder_local_synthesize"], forbiddenTools: [], booleanCriteria: [{ criterion: "Output contains meaningful structured content (not just errors or empty results)", weight: 2 }, { criterion: "Output contains entity or topic names from the query", weight: 1 }, { criterion: "At least one expected tool completed successfully", weight: 2 }] },
@@ -372,10 +372,10 @@ function operatorTemplates() {
     return [
         { query: "Show me the system health dashboard for today", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output presents system health metrics", weight: 1 }, { criterion: "Output highlights any degraded services", weight: 1 }, { criterion: "Output is operational in tone", weight: 1 }] },
         { query: "What incidents happened this week and are they resolved?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output lists incidents", weight: 1 }, { criterion: "Output includes resolution status", weight: 1 }, { criterion: "Output identifies root causes", weight: 1 }] },
-        { query: "Run a health check on all MCP infrastructure", scenario: "company_search", expectedTools: ["check_mcp_setup", "get_ops_dashboard"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output checks multiple infrastructure components", weight: 1 }, { criterion: "Output reports pass/fail per component", weight: 1 }, { criterion: "Output suggests fixes for failures", weight: 1 }] },
+        { query: "Run a health check on all MCP infrastructure", scenario: "company_search", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output checks multiple infrastructure components", weight: 1 }, { criterion: "Output reports pass/fail per component", weight: 1 }, { criterion: "Output suggests fixes for failures", weight: 1 }] },
         { query: "Delegate the on-call rotation setup to the SRE team", scenario: "delegation", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output contains delegation instructions", weight: 1 }, { criterion: "Output specifies SRE-relevant details", weight: 1 }, { criterion: "Output includes escalation paths", weight: 1 }] },
         { query: "What deployments went out this week and did any cause issues?", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output references deployments", weight: 1 }, { criterion: "Output correlates deployments with incidents", weight: 1 }, { criterion: "Output identifies rollback candidates", weight: 1 }] },
-        { query: "Compare our uptime this month vs last month", scenario: "packet_diff", expectedTools: ["founder_local_synthesize", "get_ops_dashboard"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output includes uptime percentages or trends", weight: 1 }, { criterion: "Output identifies the biggest contributor to downtime", weight: 1 }, { criterion: "Output does not fabricate exact uptime numbers", weight: 2 }] },
+        { query: "Compare our uptime this month vs last month", scenario: "packet_diff", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output includes uptime percentages or trends", weight: 1 }, { criterion: "Output identifies the biggest contributor to downtime", weight: 1 }, { criterion: "Output does not fabricate exact uptime numbers", weight: 2 }] },
         { query: "Export the incident report for the API outage", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output follows incident report structure", weight: 1 }, { criterion: "Output includes timeline, impact, and root cause", weight: 1 }, { criterion: "Output is shareable with stakeholders", weight: 1 }] },
         { query: "Flag any alerts that have been unacknowledged for over 24 hours", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output identifies stale alerts", weight: 1 }, { criterion: "Output includes age of each alert", weight: 1 }, { criterion: "Output suggests escalation for critical ones", weight: 1 }] },
         { query: "Switch to researcher mode to investigate the performance regression", scenario: "role_switch", expectedTools: ["founder_local_synthesize"], forbiddenTools: [], booleanCriteria: [{ criterion: "Output shifts to investigation perspective", weight: 1 }, { criterion: "Output suggests diagnostic tools and approaches", weight: 1 }, { criterion: "Output identifies data to collect", weight: 1 }] },
@@ -387,7 +387,7 @@ function legalTemplates() {
     return [
         { query: "Check our contracts for compliance with the new data privacy regulation", scenario: "company_search", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output references data privacy regulations", weight: 1 }, { criterion: "Output identifies compliance gaps", weight: 1 }, { criterion: "Output does not provide actual legal advice", weight: 1 }] },
         { query: "What legal risks should we flag this week?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output identifies legal risk categories", weight: 1 }, { criterion: "Output prioritizes risks by severity", weight: 1 }, { criterion: "Output includes a disclaimer about not being legal counsel", weight: 1 }] },
-        { query: "Compare the terms of our vendor contracts for consistency", scenario: "competitor_brief", expectedTools: ["compare_options", "check_contract_compliance"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output compares contract terms systematically", weight: 1 }, { criterion: "Output identifies inconsistencies", weight: 1 }, { criterion: "Output suggests standardization opportunities", weight: 1 }] },
+        { query: "Compare the terms of our vendor contracts for consistency", scenario: "competitor_brief", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output compares contract terms systematically", weight: 1 }, { criterion: "Output identifies inconsistencies", weight: 1 }, { criterion: "Output suggests standardization opportunities", weight: 1 }] },
         { query: "Export the contract review findings for outside counsel", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output is formal and counsel-appropriate", weight: 1 }, { criterion: "Output includes numbered findings", weight: 1 }, { criterion: "Output preserves legal terminology", weight: 1 }] },
         { query: "Flag any IP-related changes in our competitor filings", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output references IP or patent filings", weight: 1 }, { criterion: "Output identifies specific competitors", weight: 1 }, { criterion: "Output assesses impact on our position", weight: 1 }] },
         { query: "Prepare a delegation brief for the paralegal on discovery tasks", scenario: "delegation", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output is delegation-appropriate", weight: 1 }, { criterion: "Output specifies legal discovery requirements", weight: 1 }, { criterion: "Output includes deadlines", weight: 1 }] },
@@ -395,13 +395,13 @@ function legalTemplates() {
         { query: "Switch to banker mode to assess the financial exposure from this lawsuit", scenario: "role_switch", expectedTools: ["founder_local_synthesize"], forbiddenTools: [], booleanCriteria: [{ criterion: "Output adopts a financial assessment perspective", weight: 1 }, { criterion: "Output estimates exposure ranges", weight: 1 }, { criterion: "Output caveats financial estimates appropriately", weight: 1 }] },
         { query: "Review the NDA template for common issues", scenario: "company_search", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output references NDA-specific terms", weight: 1 }, { criterion: "Output identifies common NDA pitfalls", weight: 1 }, { criterion: "Output suggests improvements", weight: 1 }] },
         { query: "What regulatory filings are due this month?", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output lists upcoming deadlines", weight: 1 }, { criterion: "Output includes filing types", weight: 1 }, { criterion: "Output suggests preparation steps", weight: 1 }] },
-        { query: "Summarize the liability exposure across all active contracts", scenario: "company_search", expectedTools: ["check_contract_compliance", "get_recon_summary"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output addresses liability specifically", weight: 1 }, { criterion: "Output categorizes by contract type", weight: 1 }, { criterion: "Output does not fabricate liability amounts", weight: 2 }] },
+        { query: "Summarize the liability exposure across all active contracts", scenario: "company_search", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output addresses liability specifically", weight: 1 }, { criterion: "Output categorizes by contract type", weight: 1 }, { criterion: "Output does not fabricate liability amounts", weight: 2 }] },
     ];
 }
 function pmTemplates() {
     return [
         { query: "What's the status of all feature requests from this sprint?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output lists feature requests", weight: 1 }, { criterion: "Output includes status per feature", weight: 1 }, { criterion: "Output identifies blockers", weight: 1 }] },
-        { query: "Compare the user feedback for Feature A vs Feature B", scenario: "competitor_brief", expectedTools: ["compare_options"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output compares two features", weight: 1 }, { criterion: "Output references user feedback", weight: 1 }, { criterion: "Output includes a recommendation", weight: 1 }] },
+        { query: "Compare the user feedback for Feature A vs Feature B", scenario: "competitor_brief", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output compares two features", weight: 1 }, { criterion: "Output references user feedback", weight: 1 }, { criterion: "Output includes a recommendation", weight: 1 }] },
         { query: "Prepare a sprint retrospective document", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output follows retro format (what went well, what didn't, actions)", weight: 1 }, { criterion: "Output is specific to the current sprint", weight: 1 }, { criterion: "Output includes actionable improvements", weight: 1 }] },
         { query: "What user-facing changes went live this week?", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output lists specific changes", weight: 1 }, { criterion: "Output focuses on user impact", weight: 1 }, { criterion: "Output includes release dates", weight: 1 }] },
         { query: "Create a PRD outline for the new onboarding flow", scenario: "delegation", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output follows PRD structure", weight: 1 }, { criterion: "Output includes user stories or acceptance criteria", weight: 1 }, { criterion: "Output is scoped appropriately", weight: 1 }] },
@@ -420,7 +420,7 @@ function contractorTemplates() {
         { query: "Export my weekly deliverables report for the client", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output is client-facing in tone", weight: 1 }, { criterion: "Output lists deliverables with status", weight: 1 }, { criterion: "Output includes hours or effort summary", weight: 1 }] },
         { query: "What changed in the project requirements since I was last briefed?", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output identifies specific requirement changes", weight: 1 }, { criterion: "Output highlights impact on current work", weight: 1 }, { criterion: "Output suggests clarification questions", weight: 1 }] },
         { query: "Compare the scope of my current contract vs the original SOW", scenario: "packet_diff", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output compares current vs original scope", weight: 1 }, { criterion: "Output identifies scope expansion", weight: 1 }, { criterion: "Output suggests contract amendment if needed", weight: 1 }] },
-        { query: "Find the coding standards document for this project", scenario: "company_search", expectedTools: ["discover_tools", "get_project_context"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output helps locate documentation", weight: 1 }, { criterion: "Output is specific to coding standards", weight: 1 }, { criterion: "Output suggests follow-up resources", weight: 1 }] },
+        { query: "Find the coding standards document for this project", scenario: "company_search", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output helps locate documentation", weight: 1 }, { criterion: "Output is specific to coding standards", weight: 1 }, { criterion: "Output suggests follow-up resources", weight: 1 }] },
         { query: "Delegate the testing tasks to the QA contractor", scenario: "delegation", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output contains test delegation details", weight: 1 }, { criterion: "Output specifies test scope and criteria", weight: 1 }, { criterion: "Output includes acceptance standards", weight: 1 }] },
         { query: "Switch to PM mode to understand the feature priority", scenario: "role_switch", expectedTools: ["founder_local_synthesize"], forbiddenTools: [], booleanCriteria: [{ criterion: "Output adopts a PM perspective", weight: 1 }, { criterion: "Output discusses prioritization frameworks", weight: 1 }, { criterion: "Output helps contextualize current work", weight: 1 }] },
         { query: "Flag any blockers that are preventing my progress", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output identifies specific blockers", weight: 1 }, { criterion: "Output suggests workarounds or escalation paths", weight: 1 }, { criterion: "Output includes who can unblock", weight: 1 }] },
@@ -436,7 +436,7 @@ function investorTemplates() {
         { query: "What's changed in the macro environment that affects our thesis?", scenario: "important_change", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output references macroeconomic factors", weight: 1 }, { criterion: "Output connects macro to investment thesis", weight: 1 }, { criterion: "Output is data-driven, not speculative", weight: 1 }] },
         { query: "Track how our portfolio company valuations shifted this quarter", scenario: "packet_diff", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output tracks valuation changes", weight: 1 }, { criterion: "Output identifies up-rounds and down-rounds", weight: 1 }, { criterion: "Output does not fabricate specific valuations", weight: 2 }] },
         { query: "Delegate the market sizing analysis to the associate", scenario: "delegation", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_deep_context_gather"], booleanCriteria: [{ criterion: "Output includes market sizing methodology", weight: 1 }, { criterion: "Output specifies data sources to use", weight: 1 }, { criterion: "Output includes expected deliverable format", weight: 1 }] },
-        { query: "Switch to founder mode and evaluate the product from a builder's lens", scenario: "role_switch", expectedTools: ["discover_tools"], forbiddenTools: ["check_contract_compliance"], booleanCriteria: [{ criterion: "Output shifts to builder/product perspective", weight: 1 }, { criterion: "Output evaluates technical feasibility", weight: 1 }, { criterion: "Output identifies product-market fit signals", weight: 1 }] },
+        { query: "Switch to founder mode and evaluate the product from a builder's lens", scenario: "role_switch", expectedTools: ["discover_tools"], forbiddenTools: ["founder_packet_validate"], booleanCriteria: [{ criterion: "Output shifts to builder/product perspective", weight: 1 }, { criterion: "Output evaluates technical feasibility", weight: 1 }, { criterion: "Output identifies product-market fit signals", weight: 1 }] },
         { query: "Give me the weekly portfolio pulse", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output covers portfolio companies", weight: 1 }, { criterion: "Output highlights winners and at-risk companies", weight: 1 }, { criterion: "Output is concise for a weekly cadence", weight: 1 }] },
         { query: "What deal flow came in this week worth evaluating?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output references deal flow", weight: 1 }, { criterion: "Output includes basic screening criteria", weight: 1 }, { criterion: "Output recommends which to pursue", weight: 1 }] },
         { query: "Research the competitive landscape for this fintech vertical", scenario: "competitor_brief", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output maps the fintech competitive landscape", weight: 1 }, { criterion: "Output identifies market leaders and challengers", weight: 1 }, { criterion: "Output assesses white space opportunities", weight: 1 }] },
@@ -444,7 +444,7 @@ function investorTemplates() {
 }
 function contentTemplates() {
     return [
-        { query: "Draft a LinkedIn post about our latest product launch", scenario: "memo_export", expectedTools: ["founder_local_synthesize", "compress_or_expand_text"], forbiddenTools: ["run_recon", "founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output is formatted for LinkedIn", weight: 1 }, { criterion: "Output is under 300 words", weight: 1 }, { criterion: "Output includes a hook and CTA", weight: 1 }] },
+        { query: "Draft a LinkedIn post about our latest product launch", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon", "founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output is formatted for LinkedIn", weight: 1 }, { criterion: "Output is under 300 words", weight: 1 }, { criterion: "Output includes a hook and CTA", weight: 1 }] },
         { query: "What trending topics should we create content around this week?", scenario: "weekly_reset", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output identifies trending topics", weight: 1 }, { criterion: "Output connects trends to our brand", weight: 1 }, { criterion: "Output suggests specific content formats", weight: 1 }] },
         { query: "Compare our content strategy against HubSpot and Buffer", scenario: "competitor_brief", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["founder_local_weekly_reset"], booleanCriteria: [{ criterion: "Output compares content strategies", weight: 1 }, { criterion: "Output identifies what competitors do better", weight: 1 }, { criterion: "Output includes actionable takeaways", weight: 1 }] },
         { query: "Export the content calendar for next month", scenario: "memo_export", expectedTools: ["founder_local_synthesize"], forbiddenTools: ["run_recon"], booleanCriteria: [{ criterion: "Output is calendar-structured", weight: 1 }, { criterion: "Output includes content types and topics", weight: 1 }, { criterion: "Output assigns rough dates", weight: 1 }] },
@@ -723,12 +723,51 @@ async function executeQueryTools(query, allTools) {
             effectiveTools.push("founder_local_synthesize");
         }
     }
-    // 2b. Scenario-specific seeding: seed data before tools that need prior state
+    // 2b. Web enrichment: fetch live web data for scenarios that need entity-specific content
+    let webResults = [];
+    const webEnrichScenarios = ["company_search", "competitor_brief", "important_change", "delegation", "memo_export", "weekly_reset", "packet_diff", "role_switch"];
+    if (webEnrichScenarios.includes(query.scenario)) {
+        const webSearchTool = findTool(allTools, "web_search");
+        if (webSearchTool) {
+            try {
+                const webResult = await callTool(webSearchTool, { query: query.query, maxResults: 5, provider: "gemini" });
+                totalMs += webResult.ms;
+                if (webResult.ok && webResult.result) {
+                    // Extract search results from tool output
+                    const raw = webResult.result;
+                    if (Array.isArray(raw?.results)) {
+                        webResults = raw.results.slice(0, 5);
+                    }
+                    else if (Array.isArray(raw)) {
+                        webResults = raw.slice(0, 5);
+                    }
+                    else if (raw?.content) {
+                        // MCP content block format
+                        try {
+                            const text = Array.isArray(raw.content) ? raw.content.map((b) => b.text).join("") : String(raw.content);
+                            const parsed = JSON.parse(text);
+                            if (Array.isArray(parsed?.results))
+                                webResults = parsed.results.slice(0, 5);
+                            else if (Array.isArray(parsed))
+                                webResults = parsed.slice(0, 5);
+                        }
+                        catch { /* not JSON */ }
+                    }
+                    if (webResults.length > 0) {
+                        toolsFired.push("web_search");
+                        outputs["web_search"] = webResults.map(r => `${r.title}: ${r.snippet} (${r.url})`).join("\n");
+                    }
+                }
+            }
+            catch { /* web search unavailable — continue without */ }
+        }
+    }
+    // 2c. Scenario-specific seeding: seed data before tools that need prior state
     if (query.scenario === "competitor_brief") {
-        // Seed with query-specific competitor brief via founder_local_synthesize
+        // Seed with query-specific competitor brief via founder_local_synthesize + web results
         const synthTool = findTool(allTools, "founder_local_synthesize");
         if (synthTool && !toolsFired.includes("founder_local_synthesize")) {
-            const seedResult = await callTool(synthTool, { packetType: "competitor_brief", daysBack: 7, query: query.query });
+            const seedResult = await callTool(synthTool, { packetType: "competitor_brief", daysBack: 7, query: query.query, lens: query.persona, webResults });
             totalMs += seedResult.ms;
             if (seedResult.ok) {
                 toolsFired.push("founder_local_synthesize");
@@ -740,7 +779,7 @@ async function executeQueryTools(query, allTools) {
         // Seed with an important_change packet that shows what changed (our best before/after proxy)
         const synthTool = findTool(allTools, "founder_local_synthesize");
         if (synthTool && !toolsFired.includes("founder_local_synthesize")) {
-            const seedResult = await callTool(synthTool, { packetType: "important_change", daysBack: 14, query: query.query });
+            const seedResult = await callTool(synthTool, { packetType: "important_change", daysBack: 14, query: query.query, lens: query.persona, webResults });
             totalMs += seedResult.ms;
             if (seedResult.ok) {
                 toolsFired.push("founder_local_synthesize");
@@ -777,8 +816,13 @@ async function executeQueryTools(query, allTools) {
         }
         const tool = findTool(allTools, toolName);
         if (tool) {
-            // Build minimal args based on tool name patterns
+            // Build minimal args based on tool name patterns, inject webResults for synthesize
             const args = buildMinimalArgs(toolName, query);
+            if (toolName === "founder_local_synthesize") {
+                if (webResults.length > 0)
+                    args.webResults = webResults;
+                args.lens = query.persona;
+            }
             const result = await callTool(tool, args);
             totalMs += result.ms;
             if (result.ok) {
@@ -875,10 +919,18 @@ function buildMinimalArgs(toolName, query) {
             return {};
         case "check_mcp_setup":
             return {};
-        case "check_contract_compliance":
+        case "founder_packet_validate":
+            return { query: query.query };
+        case "search_all_knowledge":
+            return { query: query.query };
+        case "founder_local_weekly_reset":
+            return { daysBack: 7, query: query.query };
+        case "export_artifact_packet":
+            return { query: query.query };
+        case "render_decision_memo":
+            return { query: query.query };
+        case "start_dogfood_session":
             return { query: query.query };
-        case "build_research_digest":
-            return { topic: query.query };
         case "get_project_context":
             return {};
         case "compress_or_expand_text":