Feat : Integrate hooks in LiteLLM to modify/reject requests and responses #3959

rzeta-10 · 2026-01-14T10:06:27Z

No description provided.

…ling

devsecopsbot

🤖 AI Security analysis: "Docker Compose exposes a service port on all interfaces, allowing access from any network. This increases risk of unauthorized access, data exposure, and remote exploitation if the service lacks proper authentication and network restrictions."

Risk Level	AI Score
🟢 LOW	35.0/100

Top 4 security issues / 4 total (Critical: 0, High: 0, Medium: 4, Low: 0)

Title	Location	Recommendation
Service port exposed on all interfaces	`apps/guardrails-service/litellm/docker-compose.yaml:12`	Bind the service port to a specific IP address, e.g., '127.0.0.1:$PORT'.
Service port exposed on all interfaces	`apps/guardrails-service/litellm/docker-compose.yaml:42`	Bind the service port to a specific IP address, e.g., '127.0.0.1:$PORT'.
Service port exposed on all interfaces	`apps/guardrails-service/litellm/docker-compose.yaml:78`	Bind the service port to a specific IP address, e.g., '127.0.0.1:$PORT'.
Service port exposed on all interfaces	`apps/guardrails-service/litellm/docker-compose.yaml:95`	Bind the service port to a specific IP address, e.g., '127.0.0.1:$PORT'.

🔗 View Detailed Report

devsecopsbot · 2026-01-14T10:07:54Z

apps/guardrails-service/litellm/docker-compose.yaml

+    restart: always
+    networks:
+      - guardrails-network
+    ports:


- "127.0.0.1:80:8080"

🟡 MEDIUM: Service port exposed on all interfaces
Bind the published port to localhost to avoid exposing the service on all host interfaces. If external access is required, bind to a specific allowed IP or use a reverse proxy with proper access controls.

devsecopsbot · 2026-01-14T10:07:54Z

apps/guardrails-service/litellm/docker-compose.yaml

+    image: docker.litellm.ai/berriai/litellm:main-stable
+    networks:
+      - guardrails-network
+    ports:


- "127.0.0.1:4000:4000"

🟡 MEDIUM: Service port exposed on all interfaces
Bind the published port to localhost (127.0.0.1) so the service is not exposed on all host interfaces; this limits access to the local machine and reduces unintended external exposure.

devsecopsbot · 2026-01-14T10:07:54Z

apps/guardrails-service/litellm/docker-compose.yaml

+      POSTGRES_DB: "${POSTGRES_DB}"
+      POSTGRES_USER: "${POSTGRES_USER}"
+      POSTGRES_PASSWORD: "${POSTGRES_PASSWORD}"
+    ports:


- "127.0.0.1:5432:5432"

🟡 MEDIUM: Service port exposed on all interfaces
Bind the published PostgreSQL port to localhost to avoid exposing the database on all host interfaces (prevents external access from remote hosts). If remote access is required, use a secure network or firewall instead.

devsecopsbot · 2026-01-14T10:07:54Z

apps/guardrails-service/litellm/docker-compose.yaml

+    volumes:
+      - prometheus_data:/prometheus
+      - ./prometheus.yml:/etc/prometheus/prometheus.yml
+    ports:


- "127.0.0.1:9091:9090"

🟡 MEDIUM: Service port exposed on all interfaces
Bind Prometheus host port to localhost to avoid exposing the service on all network interfaces. This restricts access to the local machine (prevents external access).

…curity

Copilot

Pull request overview

This PR integrates LiteLLM proxy support into the Akto platform, enabling traffic capture and guardrails validation for LiteLLM-based AI agent deployments. The integration follows the existing pattern used for N8N, Langchain, and Copilot Studio connectors.

Changes:

Added LiteLLM connector type with configuration constants and validation in Java backend
Implemented Python-based custom hooks for LiteLLM proxy to intercept and validate requests via guardrails service
Created Docker Compose setup with LiteLLM, PostgreSQL, Prometheus, and guardrails service integration
Added frontend UI components and constants for LiteLLM connector configuration

Reviewed changes

Copilot reviewed 11 out of 12 changed files in this pull request and generated 16 comments.

Show a summary per file

File	Description
libs/utils/src/main/java/com/akto/jobs/executors/AIAgentConnectorConstants.java	Added LITELLM connector type constant and configuration keys
libs/utils/src/main/java/com/akto/jobs/executors/AIAgentConnectorUtils.java	Extended validation to include LITELLM connector type
apps/dashboard/src/main/java/com/akto/action/AIAgentConnectorImportAction.java	Added LiteLLM-specific parameters and configuration building logic
apps/account-job-executor/src/main/java/com/akto/account_job_executor/executor/executors/AIAgentConnectorExecutor.java	Implemented LiteLLM connector execution logic
apps/guardrails-service/litellm/custom_hooks.py	Core Python implementation of LiteLLM hooks for request validation
apps/guardrails-service/litellm/docker-compose.yaml	Multi-service Docker setup for LiteLLM with guardrails integration
apps/guardrails-service/litellm/config.yaml	LiteLLM proxy configuration with custom hooks
apps/guardrails-service/litellm/prometheus.yml	Prometheus monitoring configuration
apps/guardrails-service/litellm/.env.example	Environment variable template for LiteLLM setup
apps/dashboard/web/public/litellm.svg	LiteLLM logo asset
apps/dashboard/web/polaris_web/web/src/apps/dashboard/pages/quick_start/constants/aiAgentConnectorConstants.js	Frontend constants for LiteLLM connector
apps/dashboard/web/polaris_web/web/src/apps/dashboard/pages/quick_start/transform.js	UI integration for LiteLLM connector setup

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-01-19T06:40:51Z

apps/dashboard/src/main/java/com/akto/action/AIAgentConnectorImportAction.java

+            case CONNECTOR_TYPE_LITELLM:
+                if (litellmUrl == null || litellmUrl.isEmpty() || litellmApiKey == null
+                        || litellmApiKey.isEmpty()) {
+                    loggerMaker.error("Missing required LiteLLM configuration", LogDb.DASHBOARD);
+                    return null;
+                }
+                config.put(CONFIG_LITELLM_BASE_URL, litellmUrl);
+                config.put(CONFIG_LITELLM_API_KEY, litellmApiKey);
+                break;


The dataIngestionUrl parameter is added to the config map without any validation. If this URL is null or empty, it will still be added to the configuration, potentially causing issues downstream. While the other connector types validate their required fields, there's no validation for this common configuration parameter that appears to be mandatory across all connector types.

Copilot · 2026-01-19T06:40:52Z

apps/guardrails-service/litellm/prometheus.yml

@@ -0,0 +1,8 @@
+global:
+  scrape_interval: 15s


The Prometheus scrape interval is set to 15 seconds, which may be too aggressive for a guardrails service that could be processing high volumes of requests. This could add unnecessary load on the LiteLLM service. Consider increasing the interval to 30s or 60s unless real-time monitoring at 15-second granularity is specifically required.

Suggested change

scrape_interval: 15s

scrape_interval: 30s

Copilot · 2026-01-19T06:40:52Z

apps/guardrails-service/litellm/custom_hooks.py

+
+class GuardrailsHandler(CustomLogger):
+    def __init__(self):
+        super().__init__()
+        self.client = httpx.AsyncClient(timeout=TIMEOUT)


The httpx.AsyncClient is created without any connection limits or proper resource management in the constructor. This could lead to connection pool exhaustion under heavy load. Consider adding connection limits using httpx.Limits and implementing proper lifecycle management to ensure the client is closed when the handler is destroyed.

Suggested change

class GuardrailsHandler(CustomLogger):

def __init__(self):

super().__init__()

self.client = httpx.AsyncClient(timeout=TIMEOUT)

MAX_CONNECTIONS = int(os.getenv("GUARDRAILS_MAX_CONNECTIONS", "100"))

MAX_KEEPALIVE_CONNECTIONS = int(os.getenv("GUARDRAILS_MAX_KEEPALIVE_CONNECTIONS", "20"))

class GuardrailsHandler(CustomLogger):

def __init__(self):

super().__init__()

self.client = httpx.AsyncClient(

timeout=TIMEOUT,

limits=httpx.Limits(

max_connections=MAX_CONNECTIONS,

max_keepalive_connections=MAX_KEEPALIVE_CONNECTIONS,

),

)

Copilot · 2026-01-19T06:40:52Z

apps/guardrails-service/litellm/custom_hooks.py

+    async def _validate_background(self, data: dict, call_type: str):
+        try:
+            allowed, _ = await self._call_guardrails(data, call_type)
+            if not allowed:
+                logger.warning("Guardrails violation detected (async)")
+        except Exception as e:
+            logger.error(f"Guardrails error (async): {e}")


The error handling catches all exceptions broadly with 'except Exception', which includes system-level exceptions that should typically propagate. This makes it difficult to distinguish between expected errors (like network timeouts) and unexpected errors (like programming bugs). Consider catching specific exception types like httpx.HTTPError, httpx.TimeoutException, and json.JSONDecodeError separately for better error diagnostics and handling.

Copilot · 2026-01-19T06:40:52Z

apps/guardrails-service/litellm/custom_hooks.py

+    async def _call_guardrails(
+        self,
+        data: dict,
+        call_type: str,
+    ) -> Tuple[bool, str]:
+        if not GUARDRAILS_URL:
+            return True, ""
+
+        query = ""
+        if "messages" in data:
+            for m in data["messages"]:
+                content = m.get("content", "")
+                if isinstance(content, list):
+                    for item in content:
+                        if item.get("type") == "text":
+                            query += item.get("text", "") + " "
+                elif isinstance(content, str):
+                    query += content + " "
+        else:
+            query = data.get("prompt", "")
+
+        payload = {
+            "query": query.strip(),
+            "model": data.get("model", ""),
+        }
+
+        resp = await self.client.post(
+            f"{GUARDRAILS_URL}/api/validate/request",
+            json={
+                "payload": json.dumps(payload),
+                "call_type": call_type,
+            },
+        )
+
+        if resp.status_code != 200:
+            raise RuntimeError(f"Guardrails HTTP {resp.status_code}")
+
+        result = resp.json()
+        return (
+            result.get("Allowed", result.get("allowed", True)),
+            result.get("Reason", result.get("reason", "")),
+        )


The function returns a tuple of (bool, str) for the allowed status and reason, but the reason value is retrieved from the response and never used in the calling code. The _validate_and_block and _validate_background methods both ignore the second return value. Consider either using the reason for logging/error messages, or simplifying the return type to just return a boolean.

Copilot · 2026-01-19T06:40:54Z

apps/guardrails-service/litellm/docker-compose.yaml

+      - "--detailed_debug"
+    environment:
+      DATABASE_URL: "postgresql://${POSTGRES_USER}:${POSTGRES_PASSWORD}@db:5432/${POSTGRES_DB}"
+      STORE_MODEL_IN_DB: "True"


The docker-compose.yaml file references environment variables from a .env file, but the litellm service also needs access to GUARDRAILS_BASE_URL, SYNC_MODE, and GUARDRAILS_TIMEOUT which are defined in .env.example. However, these are not explicitly listed in the environment section of the litellm service, relying only on env_file. While this works, it would be more explicit and maintainable to list these critical environment variables in the environment section with their default values or references to make the dependencies clear.

Suggested change

STORE_MODEL_IN_DB: "True"

STORE_MODEL_IN_DB: "True"

GUARDRAILS_BASE_URL: "${GUARDRAILS_BASE_URL}"

SYNC_MODE: "${SYNC_MODE}"

GUARDRAILS_TIMEOUT: "${GUARDRAILS_TIMEOUT}"

Copilot · 2026-01-19T06:40:54Z

apps/guardrails-service/litellm/custom_hooks.py

+        if resp.status_code != 200:
+            raise RuntimeError(f"Guardrails HTTP {resp.status_code}")
+
+        result = resp.json()


The response from the guardrails service is parsed as JSON without checking the content-type header. If the service returns an error page (HTML) or plain text instead of JSON, the resp.json() call will raise an exception that gets caught in the generic exception handler, potentially masking the real error. Consider checking the response content-type or handling JSON decode errors specifically to provide better error messages.

Suggested change

result = resp.json()

content_type = resp.headers.get("content-type", "")

if "application/json" not in content_type:

# Guardrails service responded with a non-JSON payload; surface a clear error.

body_preview = resp.text[:200]

raise RuntimeError(

f"Guardrails response is not JSON (content-type={content_type!r}). "

f"Body preview: {body_preview!r}"

)

try:

result = resp.json()

except (ValueError, httpx.DecodingError) as exc:

body_preview = resp.text[:200]

raise RuntimeError(

f"Failed to decode Guardrails JSON response: {exc}. "

f"Body preview: {body_preview!r}"

) from exc

Copilot · 2026-01-19T06:40:55Z

apps/guardrails-service/litellm/docker-compose.yaml

+    volumes:
+      - postgres_data:/var/lib/postgresql/data
+    healthcheck:
+      test: [ "CMD-SHELL", "pg_isready -d litellm -U llmproxy" ]


The hardcoded database name 'litellm' and username 'llmproxy' in the healthcheck command don't match the dynamic environment variables POSTGRES_DB and POSTGRES_USER. This will cause the healthcheck to fail if different values are provided in the environment variables. The test command should use the environment variables instead.

Suggested change

test: [ "CMD-SHELL", "pg_isready -d litellm -U llmproxy" ]

test: [ "CMD-SHELL", "pg_isready -d \"$POSTGRES_DB\" -U \"$POSTGRES_USER\"" ]

Copilot · 2026-01-19T06:40:55Z

apps/guardrails-service/litellm/custom_hooks.py

+        asyncio.create_task(
+            self._validate_background(data, call_type)
+        )


The async background task created with asyncio.create_task is not being tracked or awaited. This could lead to unhandled exceptions being silently ignored and potential resource leaks. The task should be stored and properly managed, or exception handling should be added to ensure errors are logged appropriately.

Copilot · 2026-01-19T06:40:55Z

apps/guardrails-service/litellm/custom_hooks.py

+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+GUARDRAILS_URL = os.getenv("GUARDRAILS_BASE_URL")


The guardrails service URL is retrieved from the environment but there's no validation that it's a properly formatted URL or that it uses a secure protocol (HTTPS). If an invalid URL is provided, it will only fail at runtime when the first request is made. Consider validating the URL format during initialization and logging a warning if HTTP is used instead of HTTPS in production environments.

Suggested change

GUARDRAILS_URL = os.getenv("GUARDRAILS_BASE_URL")

def _validate_guardrails_url(url: str) -> str:

"""

Validate the guardrails base URL and log any potential issues.

This does not raise; it only logs, to avoid changing existing behavior.

"""

if not url:

logger.warning(

"GUARDRAILS_BASE_URL is not set; guardrails calls may fail at runtime."

)

return url

try:

parsed = httpx.URL(url)

except Exception as exc:

logger.error(

"Invalid GUARDRAILS_BASE_URL '%s': %s. Guardrails calls may fail at runtime.",

url,

exc,

)

return url

scheme = parsed.scheme

if scheme not in ("http", "https"):

logger.warning(

"GUARDRAILS_BASE_URL uses unsupported URL scheme '%s'.", scheme

)

elif scheme == "http":

env = os.getenv("ENV") or os.getenv("ENVIRONMENT") or os.getenv("PYTHON_ENV")

if env and env.lower() == "production":

logger.warning(

"GUARDRAILS_BASE_URL is using HTTP in a production environment; HTTPS is recommended."

)

return url

RAW_GUARDRAILS_URL = os.getenv("GUARDRAILS_BASE_URL")

GUARDRAILS_URL = _validate_guardrails_url(RAW_GUARDRAILS_URL)

…e query validation

…d improve query validation" This reverts commit 03bc3e0.

abhijeet-akto · 2026-01-25T03:23:09Z

apps/guardrails-service/litellm/.env.example

@@ -0,0 +1,23 @@
+# LiteLLM Configuration
+LITELLM_MASTER_KEY=sk-1234


Do not commit the key

abhijeet-akto · 2026-01-25T03:23:54Z

...src/main/java/com/akto/account_job_executor/executor/executors/AIAgentConnectorExecutor.java

+     * Downloads the LiteLLM shield binary from Azure Storage and executes it with
+     * config as env vars.
+     */
+    private void executeLiteLLMConnector(AccountJob job, Map<String, Object> config) throws Exception {


This won't be necessary for this integration

abhijeet-akto · 2026-01-25T03:26:55Z

...src/main/java/com/akto/account_job_executor/executor/executors/AIAgentConnectorExecutor.java

                executeCopilotStudioConnector(job, config);
                break;

+            case "LITELLM":


This won't be necessary

abhijeet-akto · 2026-01-25T03:27:20Z

apps/dashboard/src/main/java/com/akto/action/AIAgentConnectorImportAction.java

                config.put(CONFIG_DATAVERSE_CLIENT_SECRET, dataverseClientSecret);
                break;

+            case CONNECTOR_TYPE_LITELLM:


This won't be necessary

abhijeet-akto · 2026-01-25T03:28:12Z

.../polaris_web/web/src/apps/dashboard/pages/quick_start/constants/aiAgentConnectorConstants.js

 ];
+
+// LiteLLM Field Configuration
+export const LITELLM_FIELDS = [


Let's create the UI similar to what we have for AWS Bedrock.

abhijeet-akto · 2026-01-25T03:31:15Z

apps/guardrails-service/litellm/config.yaml

+  callbacks: [custom_hooks.proxy_handler_instance]
+  drop_params: true
+  set_verbose: false
+  request_timeout: 600


Keep the timeout very small for the callback, ideally < 5 seconds

abhijeet-akto · 2026-01-25T03:33:26Z

apps/guardrails-service/litellm/prometheus.yml

+global:
+  scrape_interval: 15s
+
+scrape_configs:


Why prometheus is needed?

abhijeet-akto · 2026-01-25T03:33:54Z

libs/utils/src/main/java/com/akto/jobs/executors/AIAgentConnectorUtils.java

        return CONNECTOR_TYPE_N8N.equals(connectorType) ||
-               CONNECTOR_TYPE_LANGCHAIN.equals(connectorType) ||
-               CONNECTOR_TYPE_COPILOT_STUDIO.equals(connectorType);
+                CONNECTOR_TYPE_LANGCHAIN.equals(connectorType) ||


This won't be needed as we are not scheduling any job.

abhijeet-akto · 2026-01-25T03:40:27Z

apps/guardrails-service/litellm/custom_hooks.py

@@ -0,0 +1,195 @@
+from litellm.integrations.custom_logger import CustomLogger


Move all the files for litellm integration to the folder
apps/mcp-endpoint-shield/litellm.

rzeta-10 added 8 commits January 12, 2026 15:50

Add LiteLLM connector support and related configurations

d5dcd22

frontend changes to add litellm connector support

2bb814a

undo auto lint changes

7788fec

undo auto lint changes

c036357

undo auto lint changes

3b27b08

use litellm hooks to pass through akto guardrails

522b22d

update GuardrailsHandler initialization and improve validation logic

da43ef0

refactor GuardrailsHandler to improve validation logic and async hand…

188fc6e

…ling

devsecopsbot bot reviewed Jan 14, 2026

View reviewed changes

update docker-compose.yaml to bind ports to localhost for improved se…

cb6e1b5

…curity

harshithb3304 requested a review from Copilot January 19, 2026 06:33

Copilot started reviewing on behalf of harshithb3304 January 19, 2026 06:33 View session

Copilot AI reviewed Jan 19, 2026

View reviewed changes

rzeta-10 added 9 commits January 19, 2026 12:24

refactor GuardrailsHandler to enhance client configuration and improv…

03bc3e0

…e query validation

Merge master branch

c57a72a

Merge master branch

c93a9ba

Revert unwanted linting changes

b3d79f5

Revert unwanted linting changes

833e8a9

Revert "refactor GuardrailsHandler to enhance client configuration an…

a746821

…d improve query validation" This reverts commit 03bc3e0.

refactor hook to match with the new api endpoint

e3a1011

undo auto lint changes

2bc1c11

frontend changes in transform.js

2896c53

abhijeet-akto requested changes Jan 25, 2026

View reviewed changes

abhijeet-akto changed the base branch from master to feat/litellm-integration January 25, 2026 03:50

abhijeet-akto merged commit ddb913c into akto-api-security:feat/litellm-integration Jan 25, 2026
8 checks passed

-      STORE_MODEL_IN_DB: "True"
+      STORE_MODEL_IN_DB: "True"
+      GUARDRAILS_BASE_URL: "${GUARDRAILS_BASE_URL}"
+      SYNC_MODE: "${SYNC_MODE}"
+      GUARDRAILS_TIMEOUT: "${GUARDRAILS_TIMEOUT}"

-        result = resp.json()
+        content_type = resp.headers.get("content-type", "")
+        if "application/json" not in content_type:
+            # Guardrails service responded with a non-JSON payload; surface a clear error.
+            body_preview = resp.text[:200]
+            raise RuntimeError(
+                f"Guardrails response is not JSON (content-type={content_type!r}). "
+                f"Body preview: {body_preview!r}"
+            )
+        try:
+            result = resp.json()
+        except (ValueError, httpx.DecodingError) as exc:
+            body_preview = resp.text[:200]
+            raise RuntimeError(
+                f"Failed to decode Guardrails JSON response: {exc}. "
+                f"Body preview: {body_preview!r}"
+            ) from exc

	test: [ "CMD-SHELL", "pg_isready -d litellm -U llmproxy" ]
	test: [ "CMD-SHELL", "pg_isready -d \"$POSTGRES_DB\" -U \"$POSTGRES_USER\"" ]

-GUARDRAILS_URL = os.getenv("GUARDRAILS_BASE_URL")
+def _validate_guardrails_url(url: str) -> str:
+    """
+    Validate the guardrails base URL and log any potential issues.
+    This does not raise; it only logs, to avoid changing existing behavior.
+    """
+    if not url:
+        logger.warning(
+            "GUARDRAILS_BASE_URL is not set; guardrails calls may fail at runtime."
+        )
+        return url
+    try:
+        parsed = httpx.URL(url)
+    except Exception as exc:
+        logger.error(
+            "Invalid GUARDRAILS_BASE_URL '%s': %s. Guardrails calls may fail at runtime.",
+            url,
+            exc,
+        )
+        return url
+    scheme = parsed.scheme
+    if scheme not in ("http", "https"):
+        logger.warning(
+            "GUARDRAILS_BASE_URL uses unsupported URL scheme '%s'.", scheme
+        )
+    elif scheme == "http":
+        env = os.getenv("ENV") or os.getenv("ENVIRONMENT") or os.getenv("PYTHON_ENV")
+        if env and env.lower() == "production":
+            logger.warning(
+                "GUARDRAILS_BASE_URL is using HTTP in a production environment; HTTPS is recommended."
+            )
+    return url
+RAW_GUARDRAILS_URL = os.getenv("GUARDRAILS_BASE_URL")
+GUARDRAILS_URL = _validate_guardrails_url(RAW_GUARDRAILS_URL)

		@@ -0,0 +1,23 @@
		# LiteLLM Configuration
		LITELLM_MASTER_KEY=sk-1234

		@@ -0,0 +1,195 @@
		from litellm.integrations.custom_logger import CustomLogger

Feat : Integrate hooks in LiteLLM to modify/reject requests and responses #3959

Feat : Integrate hooks in LiteLLM to modify/reject requests and responses #3959

Uh oh!

Conversation

rzeta-10 commented Jan 14, 2026

Uh oh!

devsecopsbot bot left a comment

Choose a reason for hiding this comment

Uh oh!

devsecopsbot bot Jan 14, 2026

Choose a reason for hiding this comment

Uh oh!

devsecopsbot bot Jan 14, 2026

Choose a reason for hiding this comment

Uh oh!

devsecopsbot bot Jan 14, 2026

Choose a reason for hiding this comment

Uh oh!

devsecopsbot bot Jan 14, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Jan 19, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 19, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 19, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 19, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 19, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 19, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 19, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 19, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 19, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 19, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone