From 12e2fbf97a1b9cbc8ef7193102e30176789b36bc Mon Sep 17 00:00:00 2001
From: ignorant05 <oussamabaccara05@gmail.com>
Date: Mon, 13 Oct 2025 19:36:59 +0100
Subject: [PATCH 01/12] Feature: mimic the quatifier integration in python sdk

---
 packages/sdk/src/qualifire.ts | 44 +++++++++++++++++++++++++++++++++++
 packages/sdk/src/types.ts     |  9 +++++++
 2 files changed, 53 insertions(+)
 create mode 100644 packages/sdk/src/qualifire.ts

diff --git a/packages/sdk/src/qualifire.ts b/packages/sdk/src/qualifire.ts
new file mode 100644
index 00000000..624ea604
--- /dev/null
+++ b/packages/sdk/src/qualifire.ts
@@ -0,0 +1,44 @@
+import { EvaluationResult, ReportSummaryRequest } from "./types";
+
+export class QualifireClient {
+  private static convertWithStructuredSummary(
+    evaluationResults: EvaluationResult,
+    request: ReportSummaryRequest
+  ): any {
+    return {
+      evaluations: evaluationResults,
+      structured: request.structuredSummary,
+      deep_test: request.deepTest,
+      start_time: request.startTime,
+      judge_model: request.judgeModel || null,
+    };
+  }
+
+  public static async reportSummaryToQualifire(
+    evaluationResults: EvaluationResult,
+    request: ReportSummaryRequest
+  ): Promise<void> {
+    console.info("Reporting summary to Qualifire");
+
+    const apiEvaluationResult = this.convertWithStructuredSummary(
+      evaluationResults,
+      request
+    );
+
+    const response = await fetch(`${request.qualifireUrl}/api/rogue/v1/report`, {
+      method: "POST",
+      headers: {
+        "Content-Type": "application/json",
+        "X-qualifire-key": request.qualifireApiKey,
+      },
+      body: JSON.stringify(apiEvaluationResult),
+    });
+
+    if (!response.ok) {
+      const errText = await response.text();
+      throw new Error(
+        `Qualifire report failed: ${response.status} ${response.statusText} - ${errText}`
+      );
+    }
+  }
+}
diff --git a/packages/sdk/src/types.ts b/packages/sdk/src/types.ts
index ccc1beaa..648382d6 100644
--- a/packages/sdk/src/types.ts
+++ b/packages/sdk/src/types.ts
@@ -118,6 +118,15 @@ export interface RogueClientConfig {
   retries?: number;
 }
 
+export interface ReportSummaryRequest {
+  qualifireUrl: string;
+  qualifireApiKey: string;
+  structuredSummary: boolean;
+  deepTest: boolean;
+  startTime: string;
+  judgeModel?: string;
+}
+
 // Event Types for WebSocket
 export type WebSocketEventType =
   | "job_update"

From 1cc20abf53f91e9b2b5a35d8fcbcd3cd99e2489d Mon Sep 17 00:00:00 2001
From: ignorant05 <oussamabaccara05@gmail.com>
Date: Tue, 14 Oct 2025 12:25:23 +0100
Subject: [PATCH 02/12] Fix: resolved errors and warnings as suggested

---
 packages/sdk/src/quatifier.ts | 80 +++++++++++++++++++++++++++++++++++
 packages/sdk/src/types.ts     | 10 ++++-
 2 files changed, 89 insertions(+), 1 deletion(-)
 create mode 100644 packages/sdk/src/quatifier.ts

diff --git a/packages/sdk/src/quatifier.ts b/packages/sdk/src/quatifier.ts
new file mode 100644
index 00000000..7a33f100
--- /dev/null
+++ b/packages/sdk/src/quatifier.ts
@@ -0,0 +1,80 @@
+import { EvaluationResult, ReportSummaryRequest, StructuredSummary } from "./types";
+
+interface QualifireReportPayload {
+  job_id: string;
+  evaluations: EvaluationResult;
+  structured: StructuredSummary;
+  deep_test: boolean;
+  start_time: string;
+  judge_model: string | null;
+}
+
+export interface QualifireClientOptions {
+  logger?: (message: string) => void;
+}
+
+export class QualifireClient {
+  private static convertWithStructuredSummary(
+    evaluationResults: EvaluationResult,
+    request: ReportSummaryRequest
+  ): QualifireReportPayload {
+    return {
+      job_id: request.job_id,
+      evaluations: evaluationResults,
+      structured: request.structuredSummary || null,
+      deep_test: request.deepTest,
+      start_time: request.startTime,
+      judge_model: request.judgeModel || null,
+    };
+  }
+
+  /**
+   * Reports evaluation summary to quatifier.
+   * 
+   * @param evaluationResults - The evaluation results to report
+   * @param request - Configuration including Qualifire URL, API key, and metadata
+   * @throws {Error} If the API request fails or returns a non-2xx status
+   * @returns A promise that resolves when the report is successfully submitted
+   */
+  public static async reportSummaryToQualifire(
+    evaluationResults: EvaluationResult,
+    request: ReportSummaryRequest,
+    options: QualifireClientOptions
+  ): Promise<void> {
+    options?.logger?.("Reporting summary to Qualifire");
+
+    const apiEvaluationResult = this.convertWithStructuredSummary(
+      evaluationResults,
+      request
+    );
+
+    const controller = new AbortController();
+    const timeoutId = setTimeout(() => controller.abort(), 3000);
+
+    try {
+      const response = await fetch(`${request.qualifireUrl}/api/evaluation/evaluate`, {
+        method: "POST",
+        headers: {
+          "Content-Type": "application/json",
+          "X-Qualifire-API-Key": request.qualifireApiKey,
+        },
+        body: JSON.stringify(apiEvaluationResult),
+        signal: controller.signal
+      });
+
+      if (!response.ok) {
+        const errText = await response.text();
+        throw new Error(
+          `Qualifire report failed: ${response.status} ${response.statusText} - ${errText}`
+        );
+      }
+      clearTimeout(timeoutId)
+    } catch (error) {
+      clearTimeout(timeoutId);
+      if (error instanceof Error && error.name === 'AbortError') {
+        throw new Error('Qualifire report timed out after 30 seconds');
+      }
+      throw error;
+    }
+  }
+}
diff --git a/packages/sdk/src/types.ts b/packages/sdk/src/types.ts
index 648382d6..e833498c 100644
--- a/packages/sdk/src/types.ts
+++ b/packages/sdk/src/types.ts
@@ -118,10 +118,18 @@ export interface RogueClientConfig {
   retries?: number;
 }
 
+export interface StructuredSummary {
+  overall_summary: string,
+  key_findings: string[],
+  recommendations: string[],
+  detailed_breakdown: object[]
+}
+
 export interface ReportSummaryRequest {
+  job_id: string,
   qualifireUrl: string;
   qualifireApiKey: string;
-  structuredSummary: boolean;
+  structuredSummary: StructuredSummary;
   deepTest: boolean;
   startTime: string;
   judgeModel?: string;

From 800f931e475eb690b09204e48120fa69644be89f Mon Sep 17 00:00:00 2001
From: ignorant05 <oussamabaccara05@gmail.com>
Date: Tue, 14 Oct 2025 19:26:40 +0100
Subject: [PATCH 03/12] Fix: minor errors and warnings

---
 packages/sdk/src/qualifire.ts | 74 ++++++++++++++++++++++++++---------
 packages/sdk/src/types.ts     |  2 +-
 2 files changed, 56 insertions(+), 20 deletions(-)

diff --git a/packages/sdk/src/qualifire.ts b/packages/sdk/src/qualifire.ts
index 624ea604..e25e2540 100644
--- a/packages/sdk/src/qualifire.ts
+++ b/packages/sdk/src/qualifire.ts
@@ -1,44 +1,80 @@
-import { EvaluationResult, ReportSummaryRequest } from "./types";
+import { EvaluationResult, ReportSummaryRequest, StructuredSummary } from "./types";
+
+interface QualifireReportPayload {
+  job_id: string;
+  evaluations: EvaluationResult;
+  structured: StructuredSummary | null;
+  deep_test: boolean;
+  start_time: string;
+  judge_model: string | null;
+}
+
+export interface QualifireClientOptions {
+  logger?: (message: string) => void;
+}
 
 export class QualifireClient {
   private static convertWithStructuredSummary(
     evaluationResults: EvaluationResult,
     request: ReportSummaryRequest
-  ): any {
+  ): QualifireReportPayload {
     return {
+      job_id: request.job_id,
       evaluations: evaluationResults,
-      structured: request.structuredSummary,
+      structured: request.structuredSummary || null,
       deep_test: request.deepTest,
       start_time: request.startTime,
       judge_model: request.judgeModel || null,
     };
   }
 
+  /**
+   * Reports evaluation summary to Qualifire.
+   * 
+   * @param evaluationResults - The evaluation results to report
+   * @param request - Configuration including Qualifire URL, API key, and metadata
+   * @throws {Error} If the API request fails or returns a non-2xx status
+   * @returns A promise that resolves when the report is successfully submitted
+   */
   public static async reportSummaryToQualifire(
     evaluationResults: EvaluationResult,
-    request: ReportSummaryRequest
+    request: ReportSummaryRequest,
+    options: QualifireClientOptions
   ): Promise<void> {
-    console.info("Reporting summary to Qualifire");
+    options?.logger?.("Reporting summary to Qualifire");
 
     const apiEvaluationResult = this.convertWithStructuredSummary(
       evaluationResults,
       request
     );
 
-    const response = await fetch(`${request.qualifireUrl}/api/rogue/v1/report`, {
-      method: "POST",
-      headers: {
-        "Content-Type": "application/json",
-        "X-qualifire-key": request.qualifireApiKey,
-      },
-      body: JSON.stringify(apiEvaluationResult),
-    });
-
-    if (!response.ok) {
-      const errText = await response.text();
-      throw new Error(
-        `Qualifire report failed: ${response.status} ${response.statusText} - ${errText}`
-      );
+    const controller = new AbortController();
+    const timeoutId = setTimeout(() => controller.abort(), 3000);
+
+    try {
+      const response = await fetch(`${request.qualifireUrl}/api/evaluation/evaluate`, {
+        method: "POST",
+        headers: {
+          "Content-Type": "application/json",
+          "X-Qualifire-API-Key": request.qualifireApiKey,
+        },
+        body: JSON.stringify(apiEvaluationResult),
+        signal: controller.signal
+      });
+
+      if (!response.ok) {
+        const errText = await response.text();
+        throw new Error(
+          `Qualifire report failed: ${response.status} ${response.statusText} - ${errText}`
+        );
+      }
+      clearTimeout(timeoutId);
+    } catch (error) {
+      clearTimeout(timeoutId);
+      if (error instanceof Error && error.name === 'AbortError') {
+        throw new Error('Qualifire report timed out after 3 seconds');
+      }
+      throw error;
     }
   }
 }
diff --git a/packages/sdk/src/types.ts b/packages/sdk/src/types.ts
index e833498c..7d21993a 100644
--- a/packages/sdk/src/types.ts
+++ b/packages/sdk/src/types.ts
@@ -129,7 +129,7 @@ export interface ReportSummaryRequest {
   job_id: string,
   qualifireUrl: string;
   qualifireApiKey: string;
-  structuredSummary: StructuredSummary;
+  structuredSummary?: StructuredSummary;
   deepTest: boolean;
   startTime: string;
   judgeModel?: string;

From eb9b8cc5025d9453ccf7172292c4b7132d0cb951 Mon Sep 17 00:00:00 2001
From: ignorant05 <oussamabaccara05@gmail.com>
Date: Tue, 14 Oct 2025 19:38:46 +0100
Subject: [PATCH 04/12] Fix: minor errors and warnings

---
 packages/sdk/src/qualifire.ts | 2 +-
 packages/sdk/src/types.ts     | 8 ++++----
 2 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/packages/sdk/src/qualifire.ts b/packages/sdk/src/qualifire.ts
index e25e2540..b81f705f 100644
--- a/packages/sdk/src/qualifire.ts
+++ b/packages/sdk/src/qualifire.ts
@@ -39,7 +39,7 @@ export class QualifireClient {
   public static async reportSummaryToQualifire(
     evaluationResults: EvaluationResult,
     request: ReportSummaryRequest,
-    options: QualifireClientOptions
+    options?: QualifireClientOptions
   ): Promise<void> {
     options?.logger?.("Reporting summary to Qualifire");
 
diff --git a/packages/sdk/src/types.ts b/packages/sdk/src/types.ts
index 7d21993a..c2f6f1e1 100644
--- a/packages/sdk/src/types.ts
+++ b/packages/sdk/src/types.ts
@@ -127,11 +127,11 @@ export interface StructuredSummary {
 
 export interface ReportSummaryRequest {
   job_id: string,
-  qualifireUrl: string;
-  qualifireApiKey: string;
+  qualifireUrl?: string;
+  qualifireApiKey?: string;
   structuredSummary?: StructuredSummary;
-  deepTest: boolean;
-  startTime: string;
+  deepTest?: boolean;
+  startTime?: string;
   judgeModel?: string;
 }
 

From c17f366637d4b8a730b95ac60de2162bd024a319 Mon Sep 17 00:00:00 2001
From: ignorant05 <oussamabaccara05@gmail.com>
Date: Wed, 15 Oct 2025 09:23:33 +0100
Subject: [PATCH 05/12] Fix: removed mispelled file quatifier.ts

---
 packages/sdk/src/quatifier.ts | 80 -----------------------------------
 1 file changed, 80 deletions(-)
 delete mode 100644 packages/sdk/src/quatifier.ts

diff --git a/packages/sdk/src/quatifier.ts b/packages/sdk/src/quatifier.ts
deleted file mode 100644
index 7a33f100..00000000
--- a/packages/sdk/src/quatifier.ts
+++ /dev/null
@@ -1,80 +0,0 @@
-import { EvaluationResult, ReportSummaryRequest, StructuredSummary } from "./types";
-
-interface QualifireReportPayload {
-  job_id: string;
-  evaluations: EvaluationResult;
-  structured: StructuredSummary;
-  deep_test: boolean;
-  start_time: string;
-  judge_model: string | null;
-}
-
-export interface QualifireClientOptions {
-  logger?: (message: string) => void;
-}
-
-export class QualifireClient {
-  private static convertWithStructuredSummary(
-    evaluationResults: EvaluationResult,
-    request: ReportSummaryRequest
-  ): QualifireReportPayload {
-    return {
-      job_id: request.job_id,
-      evaluations: evaluationResults,
-      structured: request.structuredSummary || null,
-      deep_test: request.deepTest,
-      start_time: request.startTime,
-      judge_model: request.judgeModel || null,
-    };
-  }
-
-  /**
-   * Reports evaluation summary to quatifier.
-   * 
-   * @param evaluationResults - The evaluation results to report
-   * @param request - Configuration including Qualifire URL, API key, and metadata
-   * @throws {Error} If the API request fails or returns a non-2xx status
-   * @returns A promise that resolves when the report is successfully submitted
-   */
-  public static async reportSummaryToQualifire(
-    evaluationResults: EvaluationResult,
-    request: ReportSummaryRequest,
-    options: QualifireClientOptions
-  ): Promise<void> {
-    options?.logger?.("Reporting summary to Qualifire");
-
-    const apiEvaluationResult = this.convertWithStructuredSummary(
-      evaluationResults,
-      request
-    );
-
-    const controller = new AbortController();
-    const timeoutId = setTimeout(() => controller.abort(), 3000);
-
-    try {
-      const response = await fetch(`${request.qualifireUrl}/api/evaluation/evaluate`, {
-        method: "POST",
-        headers: {
-          "Content-Type": "application/json",
-          "X-Qualifire-API-Key": request.qualifireApiKey,
-        },
-        body: JSON.stringify(apiEvaluationResult),
-        signal: controller.signal
-      });
-
-      if (!response.ok) {
-        const errText = await response.text();
-        throw new Error(
-          `Qualifire report failed: ${response.status} ${response.statusText} - ${errText}`
-        );
-      }
-      clearTimeout(timeoutId)
-    } catch (error) {
-      clearTimeout(timeoutId);
-      if (error instanceof Error && error.name === 'AbortError') {
-        throw new Error('Qualifire report timed out after 30 seconds');
-      }
-      throw error;
-    }
-  }
-}

From 37d81523a53d049a4a9d5b10651189a45a6888ed Mon Sep 17 00:00:00 2001
From: ignorant05 <oussamabaccara05@gmail.com>
Date: Wed, 15 Oct 2025 09:27:35 +0100
Subject: [PATCH 06/12] Fix: minor/potential errors

---
 packages/sdk/src/qualifire.ts | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/packages/sdk/src/qualifire.ts b/packages/sdk/src/qualifire.ts
index b81f705f..0049eb27 100644
--- a/packages/sdk/src/qualifire.ts
+++ b/packages/sdk/src/qualifire.ts
@@ -49,7 +49,7 @@ export class QualifireClient {
     );
 
     const controller = new AbortController();
-    const timeoutId = setTimeout(() => controller.abort(), 3000);
+    const timeoutId = setTimeout(() => controller.abort(), 30000);
 
     try {
       const response = await fetch(`${request.qualifireUrl}/api/evaluation/evaluate`, {
@@ -72,7 +72,7 @@ export class QualifireClient {
     } catch (error) {
       clearTimeout(timeoutId);
       if (error instanceof Error && error.name === 'AbortError') {
-        throw new Error('Qualifire report timed out after 3 seconds');
+        throw new Error('Qualifire report timed out after 30 seconds');
       }
       throw error;
     }

From d9d32737a075fc535605873fe13ebaaddf221d2e Mon Sep 17 00:00:00 2001
From: ignorant05 <oussamabaccara05@gmail.com>
Date: Sun, 19 Oct 2025 12:17:33 +0100
Subject: [PATCH 07/12] Fix: some potential issues as the cursor bot suggested

---
 packages/sdk/src/qualifire.ts | 20 ++++++++++++++++----
 1 file changed, 16 insertions(+), 4 deletions(-)

diff --git a/packages/sdk/src/qualifire.ts b/packages/sdk/src/qualifire.ts
index 0049eb27..1983d5ee 100644
--- a/packages/sdk/src/qualifire.ts
+++ b/packages/sdk/src/qualifire.ts
@@ -22,8 +22,8 @@ export class QualifireClient {
       job_id: request.job_id,
       evaluations: evaluationResults,
       structured: request.structuredSummary || null,
-      deep_test: request.deepTest,
-      start_time: request.startTime,
+      deep_test: request.deepTest ?? false,
+      start_time: request.startTime ?? new Date().toISOString(),
       judge_model: request.judgeModel || null,
     };
   }
@@ -43,6 +43,18 @@ export class QualifireClient {
   ): Promise<void> {
     options?.logger?.("Reporting summary to Qualifire");
 
+    const apiKey = request.qualifireApiKey;
+    const baseUrl = request.qualifireUrl ?? "https://api.qualifire.com";
+    const endpoint = `${baseUrl}/api/evaluation/evaluate`;
+
+    if (!apiKey) {
+      throw new Error("qualifireApiKey is required but was undefined");
+    }
+
+    if (!baseUrl || baseUrl === "undefined") {
+      throw new Error("Invalid qualifireUrl provided");
+    }
+
     const apiEvaluationResult = this.convertWithStructuredSummary(
       evaluationResults,
       request
@@ -52,11 +64,11 @@ export class QualifireClient {
     const timeoutId = setTimeout(() => controller.abort(), 30000);
 
     try {
-      const response = await fetch(`${request.qualifireUrl}/api/evaluation/evaluate`, {
+      const response = await fetch(endpoint, {
         method: "POST",
         headers: {
           "Content-Type": "application/json",
-          "X-Qualifire-API-Key": request.qualifireApiKey,
+          "X-Qualifire-API-Key": apiKey,
         },
         body: JSON.stringify(apiEvaluationResult),
         signal: controller.signal

From 673b0c378ed5e888e2f48250e87864a658f6c16e Mon Sep 17 00:00:00 2001
From: yuval-qf <yuval@qualifire.ai>
Date: Thu, 23 Oct 2025 16:47:12 +0300
Subject: [PATCH 08/12] Skip rogue sanity on fork PRs (#116)

---
 .github/workflows/rogue.yml | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

diff --git a/.github/workflows/rogue.yml b/.github/workflows/rogue.yml
index a967443c..edd50d72 100644
--- a/.github/workflows/rogue.yml
+++ b/.github/workflows/rogue.yml
@@ -1,7 +1,7 @@
 name: Rogue
 
 on:
-  pull_request_target:
+  pull_request:
   push:
     branches:
       - main
@@ -11,17 +11,11 @@ jobs:
   rogue_sanity:
     runs-on: ubuntu-latest
     timeout-minutes: 15
-    environment: rogue-sanity-ci-secrets
+    # Only run for PRs from within the same repository, not from forks
+    if: github.event_name != 'pull_request' || github.event.pull_request.head.repo.full_name == github.repository
     steps:
       - name: Checkout Repository
         uses: actions/checkout@v4
-        with:
-          # 1. Checkout the actual PR commit, not just the base branch
-          ref: ${{ github.event.pull_request.head.sha }}
-          # 2. DO NOT expose the GITHUB_TOKEN write-permissions to the untrusted code
-          # This is essential to prevent untrusted code from exfiltrating secrets
-          # by manipulating the repository itself.
-          persist-credentials: false
 
       - name: Install uv
         uses: astral-sh/setup-uv@v5

From db079ce46819a10201e1e665a14babae621cbb94 Mon Sep 17 00:00:00 2001
From: yuval-qf <yuval@qualifire.ai>
Date: Fri, 24 Oct 2025 00:19:37 +0300
Subject: [PATCH 09/12] FIRE-812 | Rogue-TUI | MCP Support (#115)

* Add protocol and transport settings to tui

* Fix mcp example command

* Fix log_evaluation kwargs
---
 .gitignore                                    |   3 +
 .vscode/launch.json                           |   8 ++
 .../tshirt_store_langgraph_mcp/__main__.py    |   2 +-
 .../tui/internal/tui/common_controller.go     |  41 ++++---
 packages/tui/internal/tui/eval_form.go        |  94 +++++++++++++--
 .../tui/internal/tui/eval_form_controller.go  |  54 +++++----
 packages/tui/internal/tui/eval_ui.go          | 114 ++++++++++++++++--
 packages/tui/internal/tui/evaluation.go       |  32 +++--
 .../tui/internal/tui/keyboard_controller.go   |  23 ++--
 rogue/evaluator_agent/base_evaluator_agent.py |   1 +
 10 files changed, 290 insertions(+), 82 deletions(-)

diff --git a/.gitignore b/.gitignore
index 91550c18..4bbf4544 100644
--- a/.gitignore
+++ b/.gitignore
@@ -90,3 +90,6 @@ dist/
 .adk/
 **/.rogue
 bin/
+
+# Go TUI
+packages/tui/cmd/rogue/__debug*
diff --git a/.vscode/launch.json b/.vscode/launch.json
index 3765647c..96d9bb38 100644
--- a/.vscode/launch.json
+++ b/.vscode/launch.json
@@ -63,6 +63,14 @@
             ],
             "envFile": "${workspaceFolder}/.env"
         },
+        {
+            "name": "Rogue TUI",
+            "type": "go",
+            "request": "launch",
+            "mode": "auto",
+            "program": "${workspaceFolder}/packages/tui/cmd/rogue/main.go",
+            "console": "integratedTerminal"
+        },
         {
             "name": "Rogue AIO",
             "type": "debugpy",
diff --git a/examples/mcp/tshirt_store_langgraph_mcp/__main__.py b/examples/mcp/tshirt_store_langgraph_mcp/__main__.py
index f0cb66a1..fcb7a894 100644
--- a/examples/mcp/tshirt_store_langgraph_mcp/__main__.py
+++ b/examples/mcp/tshirt_store_langgraph_mcp/__main__.py
@@ -15,7 +15,7 @@
     "--transport",
     "transport",
     default="streamable-http",
-    choices=["streamable-http", "sse"],
+    type=click.Choice(["streamable-http", "sse"]),
     help="Transport to use for the mcp server",
 )
 def main(host: str, port: int, transport: Literal["streamable-http", "sse"]) -> None:
diff --git a/packages/tui/internal/tui/common_controller.go b/packages/tui/internal/tui/common_controller.go
index 1d947a3c..4ad264be 100644
--- a/packages/tui/internal/tui/common_controller.go
+++ b/packages/tui/internal/tui/common_controller.go
@@ -45,19 +45,17 @@ func (m Model) handlePasteMsg(msg tea.PasteMsg) (Model, tea.Cmd) {
 		}
 
 		// Only paste into text fields (Agent URL and Judge Model)
-		if m.evalState.currentField <= 1 {
-			switch m.evalState.currentField {
-			case 0: // Agent URL
-				// Insert at cursor position
-				runes := []rune(m.evalState.AgentURL)
-				m.evalState.AgentURL = string(runes[:m.evalState.cursorPos]) + cleanText + string(runes[m.evalState.cursorPos:])
-				m.evalState.cursorPos += len([]rune(cleanText))
-			case 1: // Judge Model
-				// Insert at cursor position
-				runes := []rune(m.evalState.JudgeModel)
-				m.evalState.JudgeModel = string(runes[:m.evalState.cursorPos]) + cleanText + string(runes[m.evalState.cursorPos:])
-				m.evalState.cursorPos += len([]rune(cleanText))
-			}
+		switch m.evalState.currentField {
+		case 0: // Agent URL
+			// Insert at cursor position
+			runes := []rune(m.evalState.AgentURL)
+			m.evalState.AgentURL = string(runes[:m.evalState.cursorPos]) + cleanText + string(runes[m.evalState.cursorPos:])
+			m.evalState.cursorPos += len([]rune(cleanText))
+		case 3: // Judge Model
+			// Insert at cursor position
+			runes := []rune(m.evalState.JudgeModel)
+			m.evalState.JudgeModel = string(runes[:m.evalState.cursorPos]) + cleanText + string(runes[m.evalState.cursorPos:])
+			m.evalState.cursorPos += len([]rune(cleanText))
 		}
 		return m, nil
 	}
@@ -202,14 +200,17 @@ func (m Model) handleCommandSelectedMsg(msg components.CommandSelectedMsg) (Mode
 			// Use the configured model in provider/model format
 			judgeModel = m.config.SelectedProvider + "/" + m.config.SelectedModel
 		}
+		// TODO read agent url and protocol .rogue/user_config.json
 		m.evalState = &EvaluationViewState{
-			ServerURL:    m.config.ServerURL,
-			AgentURL:     "http://localhost:10001",
-			JudgeModel:   judgeModel,
-			ParallelRuns: 1,
-			DeepTest:     false,
-			Scenarios:    loadScenariosFromWorkdir(),
-			cursorPos:    len([]rune("http://localhost:10001")), // Set cursor to end of Agent URL
+			ServerURL:      m.config.ServerURL,
+			AgentURL:       "http://localhost:10001",
+			AgentProtocol:  ProtocolA2A,
+			AgentTransport: TransportHTTP,
+			JudgeModel:     judgeModel,
+			ParallelRuns:   1,
+			DeepTest:       false,
+			Scenarios:      loadScenariosFromWorkdir(),
+			cursorPos:      len([]rune("http://localhost:10001")), // Set cursor to end of Agent URL
 		}
 	case "configure_models":
 		// Open LLM configuration dialog
diff --git a/packages/tui/internal/tui/eval_form.go b/packages/tui/internal/tui/eval_form.go
index d52bdb11..760f96d2 100644
--- a/packages/tui/internal/tui/eval_form.go
+++ b/packages/tui/internal/tui/eval_form.go
@@ -37,8 +37,8 @@ func (m Model) renderNewEvaluation() string {
 
 	title := titleStyle.Render("🧪 New Evaluation")
 
-	// Helper function to render a field with inline label and value
-	renderField := func(fieldIndex int, label, value string) string {
+	// Helper function to render a text field with inline label and value
+	renderTextField := func(fieldIndex int, label, value string) string {
 		active := m.evalState.currentField == fieldIndex
 
 		labelStyle := lipgloss.NewStyle().
@@ -83,8 +83,86 @@ func (m Model) renderNewEvaluation() string {
 		return fieldContainer.Render(fieldContent)
 	}
 
+	// Helper function to render a dropdown field with indicators
+	renderDropdownField := func(fieldIndex int, label, value string) string {
+		active := m.evalState.currentField == fieldIndex
+
+		labelStyle := lipgloss.NewStyle().
+			Foreground(t.TextMuted()).
+			Background(t.Background()).
+			Width(20).
+			Align(lipgloss.Right)
+
+		valueStyle := lipgloss.NewStyle().
+			Foreground(t.Text()).
+			Background(t.Background()).
+			Padding(0, 1)
+
+		if active {
+			labelStyle = labelStyle.Foreground(t.Primary()).Bold(true)
+			valueStyle = valueStyle.
+				Foreground(t.Primary()).
+				Background(t.Background()).
+				Bold(true)
+			// Add dropdown indicators
+			value = "◀ " + value + " ▶"
+		}
+
+		// Create a full-width container for the field
+		fieldContainer := lipgloss.NewStyle().
+			Width(m.width-4).
+			Background(t.Background()).
+			Padding(0, 2)
+
+		fieldContent := lipgloss.JoinHorizontal(lipgloss.Left,
+			labelStyle.Render(label),
+			valueStyle.Render(value),
+		)
+
+		return fieldContainer.Render(fieldContent)
+	}
+
+	// Helper function to render a toggle field
+	renderToggleField := func(fieldIndex int, label, value string) string {
+		active := m.evalState.currentField == fieldIndex
+
+		labelStyle := lipgloss.NewStyle().
+			Foreground(t.TextMuted()).
+			Background(t.Background()).
+			Width(20).
+			Align(lipgloss.Right)
+
+		valueStyle := lipgloss.NewStyle().
+			Foreground(t.Text()).
+			Background(t.Background()).
+			Padding(0, 1)
+
+		if active {
+			labelStyle = labelStyle.Foreground(t.Primary()).Bold(true)
+			valueStyle = valueStyle.
+				Foreground(t.Primary()).
+				Background(t.Background()).
+				Bold(true)
+		}
+
+		// Create a full-width container for the field
+		fieldContainer := lipgloss.NewStyle().
+			Width(m.width-4).
+			Background(t.Background()).
+			Padding(0, 2)
+
+		fieldContent := lipgloss.JoinHorizontal(lipgloss.Left,
+			labelStyle.Render(label),
+			valueStyle.Render(value),
+		)
+
+		return fieldContainer.Render(fieldContent)
+	}
+
 	// Prepare field values
 	agent := m.evalState.AgentURL
+	protocol := string(m.evalState.AgentProtocol)
+	transport := string(m.evalState.AgentTransport)
 	judge := m.evalState.JudgeModel
 	deep := "❌"
 	if m.evalState.DeepTest {
@@ -93,7 +171,7 @@ func (m Model) renderNewEvaluation() string {
 
 	// Helper function to render the start button
 	renderStartButton := func() string {
-		active := m.evalState.currentField == 3
+		active := m.evalState.currentField == 5
 		var buttonText string
 
 		if m.evalSpinner.IsActive() {
@@ -151,9 +229,11 @@ func (m Model) renderNewEvaluation() string {
 
 	// Build the content sections
 	formSection := lipgloss.JoinVertical(lipgloss.Left,
-		renderField(0, "Agent URL:", agent),
-		renderField(1, "Judge LLM:", judge),
-		renderField(2, "Deep Test:", deep),
+		renderTextField(0, "Agent URL:", agent),
+		renderDropdownField(1, "Protocol:", protocol),
+		renderDropdownField(2, "Transport:", transport),
+		renderTextField(3, "Judge LLM:", judge),
+		renderToggleField(4, "Deep Test:", deep),
 	)
 
 	var infoLines []string
@@ -168,7 +248,7 @@ func (m Model) renderNewEvaluation() string {
 
 	buttonSection := renderStartButton()
 
-	helpText := helpStyle.Render("t Test Server   ↑/↓ switch fields   ←/→ move cursor    Space toggle   Enter activate   Esc Back")
+	helpText := helpStyle.Render("t Test Server   ↑/↓ switch fields   ←/→ move cursor/cycle dropdown    Space toggle   Enter activate   Esc Back")
 
 	// Calculate content area height (excluding title and help)
 	contentHeight := m.height - 6 // title(3) + help(1) + margins(2)
diff --git a/packages/tui/internal/tui/eval_form_controller.go b/packages/tui/internal/tui/eval_form_controller.go
index 558e274b..e5684055 100644
--- a/packages/tui/internal/tui/eval_form_controller.go
+++ b/packages/tui/internal/tui/eval_form_controller.go
@@ -26,7 +26,7 @@ func (m Model) handleEvalFormInput(msg tea.KeyMsg) (Model, tea.Cmd) {
 			switch m.evalState.currentField {
 			case 0:
 				m.evalState.cursorPos = len([]rune(m.evalState.AgentURL))
-			case 1:
+			case 3:
 				m.evalState.cursorPos = len([]rune(m.evalState.JudgeModel))
 			default:
 				m.evalState.cursorPos = 0
@@ -35,13 +35,13 @@ func (m Model) handleEvalFormInput(msg tea.KeyMsg) (Model, tea.Cmd) {
 		return m, nil
 
 	case "down":
-		if m.evalState.currentField < 3 { // Now includes start button (0-3)
+		if m.evalState.currentField < 5 { // 0-5 fields (AgentURL, Protocol, Transport, JudgeModel, DeepTest, StartButton)
 			m.evalState.currentField++
 			// Set cursor to end of field content when switching fields
 			switch m.evalState.currentField {
 			case 0:
 				m.evalState.cursorPos = len([]rune(m.evalState.AgentURL))
-			case 1:
+			case 3:
 				m.evalState.cursorPos = len([]rune(m.evalState.JudgeModel))
 			default:
 				m.evalState.cursorPos = 0
@@ -50,23 +50,31 @@ func (m Model) handleEvalFormInput(msg tea.KeyMsg) (Model, tea.Cmd) {
 		return m, nil
 
 	case "left":
-		if m.evalState.currentField <= 1 && m.evalState.cursorPos > 0 { // Text fields 0-2
-			m.evalState.cursorPos--
+		switch m.evalState.currentField {
+		case 0, 3: // Text fields: AgentURL, JudgeModel
+			if m.evalState.cursorPos > 0 {
+				m.evalState.cursorPos--
+			}
+		case 1: // Protocol dropdown
+			m.evalState.cycleProtocol(true) // cycle backwards
+		case 2: // Transport dropdown
+			m.evalState.cycleTransport(true) // cycle backwards
 		}
 		return m, nil
 
 	case "right":
-		if m.evalState.currentField <= 1 { // Text fields 0-1
-			// Get current field length to limit cursor
-			var fieldLen int
-			switch m.evalState.currentField {
-			case 0:
-				fieldLen = len(m.evalState.AgentURL)
-			case 1:
-				fieldLen = len(m.evalState.JudgeModel)
-			case 2:
-				fieldLen = len(fmt.Sprintf("%d", m.evalState.ParallelRuns))
+		switch m.evalState.currentField {
+		case 0: // AgentURL text field
+			fieldLen := len(m.evalState.AgentURL)
+			if m.evalState.cursorPos < fieldLen {
+				m.evalState.cursorPos++
 			}
+		case 1: // Protocol dropdown
+			m.evalState.cycleProtocol(false) // cycle forwards
+		case 2: // Transport dropdown
+			m.evalState.cycleTransport(false) // cycle forwards
+		case 3: // JudgeModel text field
+			fieldLen := len(m.evalState.JudgeModel)
 			if m.evalState.cursorPos < fieldLen {
 				m.evalState.cursorPos++
 			}
@@ -74,14 +82,14 @@ func (m Model) handleEvalFormInput(msg tea.KeyMsg) (Model, tea.Cmd) {
 		return m, nil
 
 	case "space":
-		if m.evalState.currentField == 2 { // DeepTest field is now index 2
+		if m.evalState.currentField == 4 { // DeepTest field is now index 2
 			m.evalState.DeepTest = !m.evalState.DeepTest
 			return m, nil
 		}
 
 	case "tab":
 		// Open LLM config dialog when on Judge Model field
-		if m.evalState.currentField == 1 { // JudgeModel field
+		if m.evalState.currentField == 3 { // JudgeModel field
 			llmDialog := components.NewLLMConfigDialog(m.config.APIKeys, m.config.SelectedProvider, m.config.SelectedModel)
 			m.llmDialog = &llmDialog
 			return m, nil
@@ -89,7 +97,7 @@ func (m Model) handleEvalFormInput(msg tea.KeyMsg) (Model, tea.Cmd) {
 
 	case "backspace":
 		// Handle backspace for text fields
-		if m.evalState.currentField <= 1 && m.evalState.cursorPos > 0 {
+		if m.evalState.currentField > 0 {
 			switch m.evalState.currentField {
 			case 0: // AgentURL
 				runes := []rune(m.evalState.AgentURL)
@@ -97,13 +105,13 @@ func (m Model) handleEvalFormInput(msg tea.KeyMsg) (Model, tea.Cmd) {
 					m.evalState.AgentURL = string(runes[:m.evalState.cursorPos-1]) + string(runes[m.evalState.cursorPos:])
 					m.evalState.cursorPos--
 				}
-			case 1: // JudgeModel
+			case 3: // JudgeModel
 				runes := []rune(m.evalState.JudgeModel)
 				if m.evalState.cursorPos <= len(runes) {
 					m.evalState.JudgeModel = string(runes[:m.evalState.cursorPos-1]) + string(runes[m.evalState.cursorPos:])
 					m.evalState.cursorPos--
 				}
-			case 2: // ParallelRuns (special handling for numbers)
+			case 6: // ParallelRuns (special handling for numbers)
 				if m.evalState.ParallelRuns >= 10 {
 					m.evalState.ParallelRuns /= 10
 					m.evalState.cursorPos--
@@ -118,17 +126,17 @@ func (m Model) handleEvalFormInput(msg tea.KeyMsg) (Model, tea.Cmd) {
 	default:
 		// insert character into text fields
 		s := msg.String()
-		if len(s) == 1 && m.evalState.currentField <= 2 { // Text fields 0-2
+		if len(s) == 1 {
 			switch m.evalState.currentField {
 			case 0: // AgentURL
 				runes := []rune(m.evalState.AgentURL)
 				m.evalState.AgentURL = string(runes[:m.evalState.cursorPos]) + s + string(runes[m.evalState.cursorPos:])
 				m.evalState.cursorPos++
-			case 1: // JudgeModel
+			case 3: // JudgeModel
 				runes := []rune(m.evalState.JudgeModel)
 				m.evalState.JudgeModel = string(runes[:m.evalState.cursorPos]) + s + string(runes[m.evalState.cursorPos:])
 				m.evalState.cursorPos++
-			case 2: // ParallelRuns (numeric only)
+			case 6: // ParallelRuns (numeric only)
 				if s[0] >= '0' && s[0] <= '9' {
 					numStr := fmt.Sprintf("%d", m.evalState.ParallelRuns)
 					runes := []rune(numStr)
diff --git a/packages/tui/internal/tui/eval_ui.go b/packages/tui/internal/tui/eval_ui.go
index 76a553be..bbadd423 100644
--- a/packages/tui/internal/tui/eval_ui.go
+++ b/packages/tui/internal/tui/eval_ui.go
@@ -7,14 +7,34 @@ import (
 	"path/filepath"
 )
 
+type Protocol string
+
+const (
+	ProtocolA2A Protocol = "a2a"
+	ProtocolMCP Protocol = "mcp"
+)
+
+type Transport string
+
+const (
+	// mcp transports
+	TransportSSE            Transport = "sse"
+	TransportStreamableHTTP Transport = "streamable_http"
+
+	// a2a transports
+	TransportHTTP Transport = "http"
+)
+
 // Minimal state for eval screens
 type EvaluationViewState struct {
-	ServerURL    string // Used from config, not editable in form
-	AgentURL     string
-	JudgeModel   string
-	ParallelRuns int
-	DeepTest     bool
-	Scenarios    []EvalScenario
+	ServerURL      string // Used from config, not editable in form
+	AgentURL       string
+	AgentProtocol  Protocol
+	AgentTransport Transport
+	JudgeModel     string
+	ParallelRuns   int
+	DeepTest       bool
+	Scenarios      []EvalScenario
 
 	// Runtime
 	Running  bool
@@ -31,7 +51,7 @@ type EvaluationViewState struct {
 	StructuredSummary StructuredSummary
 
 	// Editing state for New Evaluation
-	currentField int // 0: AgentURL, 1: JudgeModel, 2: DeepTest, 3: StartButton
+	currentField int // 0: AgentURL, 1: Protocol, 2: Transport, 3: JudgeModel, 4: DeepTest, 5: StartButton
 	cursorPos    int // rune index in current text field
 }
 
@@ -74,7 +94,7 @@ func loadScenariosFromWorkdir() []EvalScenario {
 
 // startEval kicks off evaluation and consumes events into state
 func (m *Model) startEval(ctx context.Context, st *EvaluationViewState) {
-	ch, cancel, err := m.StartEvaluation(ctx, st.ServerURL, st.AgentURL, st.Scenarios, st.JudgeModel, st.ParallelRuns, st.DeepTest)
+	ch, cancel, err := m.StartEvaluation(ctx, st.ServerURL, st.AgentURL, st.AgentProtocol, st.AgentTransport, st.Scenarios, st.JudgeModel, st.ParallelRuns, st.DeepTest)
 	if err != nil {
 		st.Running = false
 		st.Status = "error"
@@ -123,3 +143,81 @@ func (m *Model) triggerSummaryGeneration() {
 	// Start spinner for automatic summary generation
 	m.summarySpinner.SetActive(true)
 }
+
+// getAllProtocols returns all available protocol options
+func getAllProtocols() []Protocol {
+	return []Protocol{ProtocolA2A, ProtocolMCP}
+}
+
+// getTransportsForProtocol returns valid transport options for a given protocol
+func getTransportsForProtocol(protocol Protocol) []Transport {
+	switch protocol {
+	case ProtocolMCP:
+		return []Transport{TransportStreamableHTTP, TransportSSE}
+	case ProtocolA2A:
+		return []Transport{TransportHTTP}
+	default:
+		return []Transport{}
+	}
+}
+
+// cycleProtocol cycles to the next protocol option
+func (st *EvaluationViewState) cycleProtocol(reverse bool) {
+	protocols := getAllProtocols()
+	currentIdx := -1
+	for i, p := range protocols {
+		if p == st.AgentProtocol {
+			currentIdx = i
+			break
+		}
+	}
+
+	if reverse {
+		currentIdx--
+		if currentIdx < 0 {
+			currentIdx = len(protocols) - 1
+		}
+	} else {
+		currentIdx++
+		if currentIdx >= len(protocols) {
+			currentIdx = 0
+		}
+	}
+
+	st.AgentProtocol = protocols[currentIdx]
+	// Reset transport to first valid option for new protocol
+	validTransports := getTransportsForProtocol(st.AgentProtocol)
+	if len(validTransports) > 0 {
+		st.AgentTransport = validTransports[0]
+	}
+}
+
+// cycleTransport cycles to the next transport option for the current protocol
+func (st *EvaluationViewState) cycleTransport(reverse bool) {
+	transports := getTransportsForProtocol(st.AgentProtocol)
+	if len(transports) == 0 {
+		return
+	}
+
+	currentIdx := -1
+	for i, t := range transports {
+		if t == st.AgentTransport {
+			currentIdx = i
+			break
+		}
+	}
+
+	if reverse {
+		currentIdx--
+		if currentIdx < 0 {
+			currentIdx = len(transports) - 1
+		}
+	} else {
+		currentIdx++
+		if currentIdx >= len(transports) {
+			currentIdx = 0
+		}
+	}
+
+	st.AgentTransport = transports[currentIdx]
+}
diff --git a/packages/tui/internal/tui/evaluation.go b/packages/tui/internal/tui/evaluation.go
index 6b021191..952b62eb 100644
--- a/packages/tui/internal/tui/evaluation.go
+++ b/packages/tui/internal/tui/evaluation.go
@@ -32,13 +32,15 @@ const (
 )
 
 type AgentConfig struct {
-	EvaluatedAgentURL         string   `json:"evaluated_agent_url"`
-	EvaluatedAgentAuthType    AuthType `json:"evaluated_agent_auth_type"`
-	EvaluatedAgentCredentials string   `json:"evaluated_agent_credentials,omitempty"`
-	JudgeLLMModel             string   `json:"judge_llm"`
-	InterviewMode             bool     `json:"interview_mode"`
-	DeepTestMode              bool     `json:"deep_test_mode"`
-	ParallelRuns              int      `json:"parallel_runs"`
+	EvaluatedAgentURL         string    `json:"evaluated_agent_url"`
+	EvaluatedAgentProtocol    Protocol  `json:"protocol"`
+	EvaluatedAgentTransport   Transport `json:"transport"`
+	EvaluatedAgentAuthType    AuthType  `json:"evaluated_agent_auth_type"`
+	EvaluatedAgentCredentials string    `json:"evaluated_agent_credentials,omitempty"`
+	JudgeLLMModel             string    `json:"judge_llm"`
+	InterviewMode             bool      `json:"interview_mode"`
+	DeepTestMode              bool      `json:"deep_test_mode"`
+	ParallelRuns              int       `json:"parallel_runs"`
 }
 
 type EvalScenario struct {
@@ -421,6 +423,8 @@ func (m *Model) StartEvaluation(
 	ctx context.Context,
 	serverURL string,
 	agentURL string,
+	agentProtocol Protocol,
+	agentTransport Transport,
 	scenarios []EvalScenario,
 	judgeModel string,
 	parallelRuns int,
@@ -439,12 +443,14 @@ func (m *Model) StartEvaluation(
 	// Build evaluation request
 	request := EvaluationRequest{
 		AgentConfig: AgentConfig{
-			EvaluatedAgentURL:      agentURL,
-			EvaluatedAgentAuthType: AuthTypeNoAuth,
-			JudgeLLMModel:          judgeModel,
-			InterviewMode:          true,
-			DeepTestMode:           deepTest,
-			ParallelRuns:           parallelRuns,
+			EvaluatedAgentURL:       agentURL,
+			EvaluatedAgentProtocol:  agentProtocol,
+			EvaluatedAgentTransport: agentTransport,
+			EvaluatedAgentAuthType:  AuthTypeNoAuth,
+			JudgeLLMModel:           judgeModel,
+			InterviewMode:           true,
+			DeepTestMode:            deepTest,
+			ParallelRuns:            parallelRuns,
 		},
 		MaxRetries:     3,
 		TimeoutSeconds: 600,
diff --git a/packages/tui/internal/tui/keyboard_controller.go b/packages/tui/internal/tui/keyboard_controller.go
index 41d5528b..c38b93a9 100644
--- a/packages/tui/internal/tui/keyboard_controller.go
+++ b/packages/tui/internal/tui/keyboard_controller.go
@@ -77,14 +77,17 @@ func (m Model) handleGlobalCtrlN() (Model, tea.Cmd) {
 		// Use the configured model in provider/model format
 		judgeModel = m.config.SelectedProvider + "/" + m.config.SelectedModel
 	}
+	// TODO read agent url and protocol .rogue/user_config.json
 	m.evalState = &EvaluationViewState{
-		ServerURL:    m.config.ServerURL,
-		AgentURL:     "http://localhost:10001",
-		JudgeModel:   judgeModel,
-		ParallelRuns: 1,
-		DeepTest:     false,
-		Scenarios:    loadScenariosFromWorkdir(),
-		cursorPos:    len([]rune("http://localhost:10001")), // Set cursor to end of Agent URL
+		ServerURL:      m.config.ServerURL,
+		AgentURL:       "http://localhost:10001",
+		AgentProtocol:  ProtocolA2A,
+		AgentTransport: TransportHTTP,
+		JudgeModel:     judgeModel,
+		ParallelRuns:   1,
+		DeepTest:       false,
+		Scenarios:      loadScenariosFromWorkdir(),
+		cursorPos:      len([]rune("http://localhost:10001")), // Set cursor to end of Agent URL
 	}
 	m.currentScreen = NewEvaluationScreen
 	return m, nil
@@ -139,7 +142,7 @@ func (m Model) handleGlobalSlash(msg tea.KeyMsg) (Model, tea.Cmd) {
 			runes := []rune(m.evalState.AgentURL)
 			m.evalState.AgentURL = string(runes[:m.evalState.cursorPos]) + s + string(runes[m.evalState.cursorPos:])
 			m.evalState.cursorPos++
-		case 1: // JudgeModel
+		case 3: // JudgeModel
 			runes := []rune(m.evalState.JudgeModel)
 			m.evalState.JudgeModel = string(runes[:m.evalState.cursorPos]) + s + string(runes[m.evalState.cursorPos:])
 			m.evalState.cursorPos++
@@ -266,11 +269,11 @@ func (m Model) handleGlobalEnter(msg tea.KeyMsg) (Model, tea.Cmd) {
 	}
 	// Handle NewEvaluationScreen enter for start button and LLM config
 	if m.currentScreen == NewEvaluationScreen && m.evalState != nil {
-		if m.evalState.currentField == 3 { // Start button field
+		if m.evalState.currentField == 5 { // Start button field
 			m.handleNewEvalEnter()
 			// Return command to start evaluation after showing spinner
 			return m, tea.Batch(m.evalSpinner.Start(), startEvaluationCmd())
-		} else if m.evalState.currentField == 1 { // Judge LLM field
+		} else if m.evalState.currentField == 3 { // Judge LLM field
 			// Open LLM config dialog when Enter is pressed on Judge LLM field
 			llmDialog := components.NewLLMConfigDialog(m.config.APIKeys, m.config.SelectedProvider, m.config.SelectedModel)
 			m.llmDialog = &llmDialog
diff --git a/rogue/evaluator_agent/base_evaluator_agent.py b/rogue/evaluator_agent/base_evaluator_agent.py
index 86f1d09e..4afdc5a0 100644
--- a/rogue/evaluator_agent/base_evaluator_agent.py
+++ b/rogue/evaluator_agent/base_evaluator_agent.py
@@ -366,6 +366,7 @@ def _log_evaluation(
         evaluation_passed: bool,
         reason: str,
         scenario_type: Optional[str],
+        **kwargs,
     ) -> None:
         """
         Logs the evaluation of the given scenario and test case.

From db44d1d62b43e6f4227c6ae58844348ab1e68cd2 Mon Sep 17 00:00:00 2001
From: ignorant05 <oussamabaccara05@gmail.com>
Date: Fri, 24 Oct 2025 13:47:50 +0100
Subject: [PATCH 10/12] Fix: update path

---
 packages/sdk/src/qualifire.ts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/packages/sdk/src/qualifire.ts b/packages/sdk/src/qualifire.ts
index 1983d5ee..a57b39b5 100644
--- a/packages/sdk/src/qualifire.ts
+++ b/packages/sdk/src/qualifire.ts
@@ -45,7 +45,7 @@ export class QualifireClient {
 
     const apiKey = request.qualifireApiKey;
     const baseUrl = request.qualifireUrl ?? "https://api.qualifire.com";
-    const endpoint = `${baseUrl}/api/evaluation/evaluate`;
+    const endpoint = `${baseUrl}/api/v1/evaluations`;
 
     if (!apiKey) {
       throw new Error("qualifireApiKey is required but was undefined");

From 701181b009e59806d1c0505872fe9bf4a9fde696 Mon Sep 17 00:00:00 2001
From: ignorant05 <oussamabaccara05@gmail.com>
Date: Fri, 24 Oct 2025 14:13:09 +0100
Subject: [PATCH 11/12] Fix: wrong path

---
 packages/sdk/src/qualifire.ts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/packages/sdk/src/qualifire.ts b/packages/sdk/src/qualifire.ts
index a57b39b5..9cd4f6b6 100644
--- a/packages/sdk/src/qualifire.ts
+++ b/packages/sdk/src/qualifire.ts
@@ -45,7 +45,7 @@ export class QualifireClient {
 
     const apiKey = request.qualifireApiKey;
     const baseUrl = request.qualifireUrl ?? "https://api.qualifire.com";
-    const endpoint = `${baseUrl}/api/v1/evaluations`;
+    const endpoint = `${baseUrl}/llm/summary`;
 
     if (!apiKey) {
       throw new Error("qualifireApiKey is required but was undefined");

From edc73fe4806a39e6ef827478b7e35920a6d7260a Mon Sep 17 00:00:00 2001
From: ignorant05 <oussamabaccara05@gmail.com>
Date: Tue, 11 Nov 2025 14:16:13 +0100
Subject: [PATCH 12/12] Rmoved duplicates

---
 .../internal/screens/evaluations/form_view.go | 86 ++-----------------
 packages/tui/internal/tui/utils.go            |  5 --
 2 files changed, 5 insertions(+), 86 deletions(-)

diff --git a/packages/tui/internal/screens/evaluations/form_view.go b/packages/tui/internal/screens/evaluations/form_view.go
index 14d5dbda..4062163a 100644
--- a/packages/tui/internal/screens/evaluations/form_view.go
+++ b/packages/tui/internal/screens/evaluations/form_view.go
@@ -159,87 +159,11 @@ func RenderForm(state *FormState) string {
 		return fieldContainer.Render(fieldContent)
 	}
 
-	// Helper function to render a dropdown field with indicators
-	renderDropdownField := func(fieldIndex int, label, value string) string {
-		active := m.evalState.currentField == fieldIndex
-
-		labelStyle := lipgloss.NewStyle().
-			Foreground(t.TextMuted()).
-			Background(t.Background()).
-			Width(20).
-			Align(lipgloss.Right)
-
-		valueStyle := lipgloss.NewStyle().
-			Foreground(t.Text()).
-			Background(t.Background()).
-			Padding(0, 1)
-
-		if active {
-			labelStyle = labelStyle.Foreground(t.Primary()).Bold(true)
-			valueStyle = valueStyle.
-				Foreground(t.Primary()).
-				Background(t.Background()).
-				Bold(true)
-			// Add dropdown indicators
-			value = "◀ " + value + " ▶"
-		}
-
-		// Create a full-width container for the field
-		fieldContainer := lipgloss.NewStyle().
-			Width(m.width-4).
-			Background(t.Background()).
-			Padding(0, 2)
-
-		fieldContent := lipgloss.JoinHorizontal(lipgloss.Left,
-			labelStyle.Render(label),
-			valueStyle.Render(value),
-		)
-
-		return fieldContainer.Render(fieldContent)
-	}
-
-	// Helper function to render a toggle field
-	renderToggleField := func(fieldIndex int, label, value string) string {
-		active := m.evalState.currentField == fieldIndex
-
-		labelStyle := lipgloss.NewStyle().
-			Foreground(t.TextMuted()).
-			Background(t.Background()).
-			Width(20).
-			Align(lipgloss.Right)
-
-		valueStyle := lipgloss.NewStyle().
-			Foreground(t.Text()).
-			Background(t.Background()).
-			Padding(0, 1)
-
-		if active {
-			labelStyle = labelStyle.Foreground(t.Primary()).Bold(true)
-			valueStyle = valueStyle.
-				Foreground(t.Primary()).
-				Background(t.Background()).
-				Bold(true)
-		}
-
-		// Create a full-width container for the field
-		fieldContainer := lipgloss.NewStyle().
-			Width(m.width-4).
-			Background(t.Background()).
-			Padding(0, 2)
-
-		fieldContent := lipgloss.JoinHorizontal(lipgloss.Left,
-			labelStyle.Render(label),
-			valueStyle.Render(value),
-		)
-
-		return fieldContainer.Render(fieldContent)
-	}
-
 	// Prepare field values
-	agent := m.evalState.AgentURL
-	protocol := string(m.evalState.AgentProtocol)
-	transport := string(m.evalState.AgentTransport)
-	judge := m.evalState.JudgeModel
+	agent := state.AgentURL
+	protocol := string(state.Protocol)
+	transport := string(state.Transport)
+	judge := state.JudgeModel
 	deep := "❌"
 	if state.DeepTest {
 		deep = "✅"
@@ -247,7 +171,7 @@ func RenderForm(state *FormState) string {
 
 	// Helper function to render the start button
 	renderStartButton := func() string {
-		active := m.evalState.currentField == 5
+		active := state.CurrentField == 5
 		var buttonText string
 
 		if state.EvalSpinnerActive {
diff --git a/packages/tui/internal/tui/utils.go b/packages/tui/internal/tui/utils.go
index a19b7282..f43a5c98 100644
--- a/packages/tui/internal/tui/utils.go
+++ b/packages/tui/internal/tui/utils.go
@@ -41,11 +41,6 @@ type AgentConfig struct {
 	InterviewMode             bool      `json:"interview_mode"`
 	DeepTestMode              bool      `json:"deep_test_mode"`
 	ParallelRuns              int       `json:"parallel_runs"`
-
-type EvalScenario struct {
-	Scenario        string       `json:"scenario"`
-	ScenarioType    ScenarioType `json:"scenario_type"`
-	ExpectedOutcome string       `json:"expected_outcome,omitempty"`
 }
 
 type EvaluationRequest struct {