mirror of
https://github.com/prowler-cloud/prowler.git
synced 2026-07-24 04:51:51 +00:00
chore(aws): enhance metadata for bedrock service (#8827)
Co-authored-by: Daniel Barranquero <danielbo2001@gmail.com>
This commit is contained in:
Rubén De la Torre Vico
committed by
GitHub
GitHub
parent
e5b86da6e5
commit
0472eb74d2
@@ -42,6 +42,7 @@ All notable changes to the **Prowler SDK** are documented in this file.
|
||||
- Update AWS S3 service metadata to new format [(#9552)](https://github.com/prowler-cloud/prowler/pull/9552)
|
||||
- Update AWS DataSync service metadata to new format [(#8854)](https://github.com/prowler-cloud/prowler/pull/8854)
|
||||
- Update AWS RDS service metadata to new format [(#9551)](https://github.com/prowler-cloud/prowler/pull/9551)
|
||||
- Update AWS Bedrock service metadata to new format [(#8827)](https://github.com/prowler-cloud/prowler/pull/8827)
|
||||
|
||||
---
|
||||
|
||||
|
||||
+21
-12
@@ -1,27 +1,36 @@
|
||||
{
|
||||
"Provider": "aws",
|
||||
"CheckID": "bedrock_agent_guardrail_enabled",
|
||||
"CheckTitle": "Ensure that Guardrails are enabled for Amazon Bedrock agent sessions.",
|
||||
"CheckType": [],
|
||||
"CheckTitle": "Amazon Bedrock agent uses a guardrail to protect agent sessions",
|
||||
"CheckType": [
|
||||
"Software and Configuration Checks/AWS Security Best Practices/Runtime Behavior Analysis",
|
||||
"Effects/Data Exposure"
|
||||
],
|
||||
"ServiceName": "bedrock",
|
||||
"SubServiceName": "",
|
||||
"ResourceIdTemplate": "arn:partition:bedrock:region:account-id:agent/resource-id",
|
||||
"ResourceIdTemplate": "",
|
||||
"Severity": "high",
|
||||
"ResourceType": "Other",
|
||||
"ResourceGroup": "ai_ml",
|
||||
"Description": "This check ensures that Guardrails are enabled to protect Amazon Bedrock agent sessions. Guardrails help mitigate security risks by filtering and blocking harmful or sensitive content during interactions with AI models.",
|
||||
"Risk": "Without guardrails enabled, Amazon Bedrock agent sessions are vulnerable to harmful prompts or inputs that could expose sensitive information or generate inappropriate content. This could lead to privacy violations, data leaks, or other security risks.",
|
||||
"RelatedUrl": "https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails.html",
|
||||
"Description": "**Bedrock agents** should have an associated **guardrail** for their sessions. The evaluation identifies agents without a guardrail linked for input/output screening during interactions.",
|
||||
"Risk": "Without **guardrails**, agent exchanges may expose **PII** or internal data (confidentiality), accept **prompt injections** that manipulate tool calls or outputs (integrity), and produce unsafe or out-of-scope responses that erode trust and cause policy violations.",
|
||||
"RelatedUrl": "",
|
||||
"AdditionalURLs": [
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails-create.html",
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/agents-guardrail.html",
|
||||
"https://www.trendmicro.com/cloudoneconformity/knowledge-base/aws/Bedrock/protect-agent-sessions-with-guardrails.html",
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails.html"
|
||||
],
|
||||
"Remediation": {
|
||||
"Code": {
|
||||
"CLI": "",
|
||||
"NativeIaC": "",
|
||||
"Other": "https://www.trendmicro.com/cloudoneconformity/knowledge-base/aws/Bedrock/protect-agent-sessions-with-guardrails.html",
|
||||
"Terraform": ""
|
||||
"CLI": "aws bedrock-agent update-agent --agent-id <agent_id> --agent-name <agent_name> --agent-resource-role-arn <role_arn> --foundation-model <model_id> --guardrail-configuration guardrailIdentifier=<guardrail_id>,guardrailVersion=DRAFT",
|
||||
"NativeIaC": "```yaml\n# CloudFormation: Associate a guardrail with a Bedrock Agent\nResources:\n ExampleAgent:\n Type: AWS::Bedrock::Agent\n Properties:\n AgentName: <example_resource_name>\n AgentResourceRoleArn: <example_resource_id>\n FoundationModel: <example_resource_name>\n Instruction: \"<example_resource_name>\"\n GuardrailConfiguration: # CRITICAL: associates a guardrail to protect sessions\n GuardrailIdentifier: <example_resource_id> # CRITICAL: guardrail ID used by the agent\n GuardrailVersion: DRAFT # CRITICAL: version applied\n```",
|
||||
"Other": "1. Open the Amazon Bedrock console and go to Agents\n2. Select the agent and click Edit\n3. In Guardrail details, select an existing guardrail and its version (e.g., DRAFT)\n4. Click Save (deploy changes if prompted)\n5. Verify the agent now shows the selected guardrail",
|
||||
"Terraform": "```hcl\n# Terraform (AWS Cloud Control): Associate a guardrail with a Bedrock Agent\nresource \"awscc_bedrock_agent\" \"example\" {\n agent_name = \"<example_resource_name>\"\n agent_resource_role_arn = \"<example_resource_id>\"\n foundation_model = \"<example_resource_name>\"\n instruction = \"<example_resource_name>\"\n\n # CRITICAL: associates a guardrail to protect agent sessions\n guardrail_configuration {\n guardrail_identifier = \"<example_resource_id>\" # CRITICAL: guardrail ID\n guardrail_version = \"DRAFT\" # CRITICAL: version applied\n }\n}\n```"
|
||||
},
|
||||
"Recommendation": {
|
||||
"Text": "Enable Guardrails for Amazon Bedrock agent sessions to protect against harmful inputs and outputs during interactions.",
|
||||
"Url": "https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails-create.html"
|
||||
"Text": "Associate a **guardrail** with every agent and tailor policies to your use case:\n- Enable content/word filters, denied topics, and sensitive-data masking\n- Use contextual grounding for RAG where relevant\n- Test and iterate across versions\nApply **least privilege** to agent tools and use **defense in depth** with monitoring and review.",
|
||||
"Url": "https://hub.prowler.com/check/bedrock_agent_guardrail_enabled"
|
||||
}
|
||||
},
|
||||
"Categories": [
|
||||
|
||||
+21
-14
@@ -1,35 +1,42 @@
|
||||
{
|
||||
"Provider": "aws",
|
||||
"CheckID": "bedrock_api_key_no_administrative_privileges",
|
||||
"CheckTitle": "Ensure Amazon Bedrock API keys do not have administrative privileges or privilege escalation",
|
||||
"CheckTitle": "Amazon Bedrock API key does not have administrative privileges, privilege escalation paths, or full Bedrock service access",
|
||||
"CheckType": [
|
||||
"Software and Configuration Checks",
|
||||
"Industry and Regulatory Standards"
|
||||
"Software and Configuration Checks/AWS Security Best Practices",
|
||||
"Software and Configuration Checks/Industry and Regulatory Standards/AWS Foundational Security Best Practices",
|
||||
"Software and Configuration Checks/Industry and Regulatory Standards/CIS AWS Foundations Benchmark",
|
||||
"TTPs/Privilege Escalation"
|
||||
],
|
||||
"ServiceName": "bedrock",
|
||||
"SubServiceName": "",
|
||||
"ResourceIdTemplate": "arn:partition:iam:region:account-id:user/{user-name}/credential/{api-key-id}",
|
||||
"ResourceIdTemplate": "",
|
||||
"Severity": "high",
|
||||
"ResourceType": "AwsIamServiceSpecificCredential",
|
||||
"ResourceType": "AwsIamAccessKey",
|
||||
"ResourceGroup": "IAM",
|
||||
"Description": "Ensure that Amazon Bedrock API keys do not have administrative privileges or privilege escalation capabilities. API keys with administrative privileges can perform any action on any resource in your AWS environment, while privilege escalation allows users to grant themselves additional permissions, both posing significant security risks.",
|
||||
"Risk": "Amazon Bedrock API keys with administrative privileges can perform any action on any resource in your AWS environment. Privilege escalation capabilities allow users to grant themselves additional permissions beyond their intended scope. Both violations of the principle of least privilege can lead to security vulnerabilities, data leaks, data loss, or unexpected charges if the API key is compromised or misused.",
|
||||
"RelatedUrl": "https://docs.aws.amazon.com/bedrock/latest/userguide/api-keys.html",
|
||||
"Description": "**Bedrock API keys** linked to IAM users are evaluated for excessive permissions, including policies that grant full access (`*` or `bedrock:*`) or enable **privilege escalation**. The finding highlights keys whose attached or inline policies provide broad or escalating capabilities.",
|
||||
"Risk": "Over-privileged Bedrock API keys weaken confidentiality, integrity, and availability. If compromised, an attacker could:\n- Escalate IAM rights and persist access\n- Invoke models at scale to exfiltrate data or incur high costs\n- Modify Bedrock settings, disrupting operations",
|
||||
"RelatedUrl": "",
|
||||
"AdditionalURLs": [
|
||||
"https://docs.aws.amazon.com/IAM/latest/UserGuide/best-practices.html#grant-least-privilege",
|
||||
"https://docs.aws.amazon.com/IAM/latest/UserGuide/getting-started-reduce-permissions.html",
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/api-keys.html"
|
||||
],
|
||||
"Remediation": {
|
||||
"Code": {
|
||||
"CLI": "aws iam delete-service-specific-credential --user-name <username> --service-specific-credential-id <credential-id>",
|
||||
"NativeIaC": "",
|
||||
"Other": "",
|
||||
"Terraform": ""
|
||||
"NativeIaC": "```yaml\n# CloudFormation: attach least-privilege policy to the IAM user owning the Bedrock API key\nResources:\n <example_resource_name>:\n Type: AWS::IAM::Policy\n Properties:\n PolicyName: least-priv-bedrock\n Users:\n - <example_resource_name>\n PolicyDocument:\n Version: '2012-10-17'\n Statement:\n - Effect: Allow\n Action:\n - bedrock:InvokeModel # CRITICAL: allow only needed Bedrock action to avoid admin or bedrock:* permissions\n Resource: \"*\" # Limits access scope to only InvokeModel on any resource\n```",
|
||||
"Other": "1. Open the AWS Console and go to IAM > Users\n2. Select the user that owns the Bedrock service-specific credential (Security credentials > Service-specific credentials shows bedrock.amazonaws.com)\n3. In the Permissions tab, detach any policy granting AdministratorAccess or bedrock:* (e.g., AmazonBedrockFullAccess)\n4. In the same tab, delete any inline policy that provides admin/privilege-escalation permissions or bedrock:* access\n5. If Bedrock access is needed, add a minimal policy allowing only bedrock:InvokeModel\n6. Save changes",
|
||||
"Terraform": "```hcl\n# Attach a minimal inline policy to the IAM user owning the Bedrock API key\nresource \"aws_iam_user_policy\" \"<example_resource_name>\" {\n name = \"least-priv-bedrock\"\n user = \"<example_resource_name>\"\n\n # CRITICAL: allow only the specific action required; avoids admin or bedrock:* full access\n policy = jsonencode({\n Version = \"2012-10-17\"\n Statement = [{\n Effect = \"Allow\"\n Action = [\"bedrock:InvokeModel\"]\n Resource = \"*\"\n }]\n })\n}\n```"
|
||||
},
|
||||
"Recommendation": {
|
||||
"Text": "Apply the principle of least privilege to Amazon Bedrock API keys. Instead of granting administrative privileges or privilege escalation capabilities, assign only the permissions necessary for specific tasks. Create custom IAM policies with minimal permissions based on the principle of least privilege. Regularly review and audit API key permissions to ensure they cannot be used for privilege escalation.",
|
||||
"Url": "https://docs.aws.amazon.com/IAM/latest/UserGuide/best-practices.html#grant-least-privilege"
|
||||
"Text": "Enforce **least privilege** on Bedrock keys:\n- Avoid wildcards like `*` and `bedrock:*`; allow only required actions\n- Prevent identity changes by disallowing `iam:*`\n- Prefer short-term credentials with rotation and MFA\n- Use permissions boundaries and SCPs as guardrails\n- Review usage and tighten policies via access analysis",
|
||||
"Url": "https://hub.prowler.com/check/bedrock_api_key_no_administrative_privileges"
|
||||
}
|
||||
},
|
||||
"Categories": [
|
||||
"gen-ai",
|
||||
"trust-boundaries"
|
||||
"identity-access"
|
||||
],
|
||||
"DependsOn": [],
|
||||
"RelatedTo": [],
|
||||
|
||||
+20
-13
@@ -1,35 +1,42 @@
|
||||
{
|
||||
"Provider": "aws",
|
||||
"CheckID": "bedrock_api_key_no_long_term_credentials",
|
||||
"CheckTitle": "Ensure Amazon Bedrock API keys are not long-term credentials",
|
||||
"CheckTitle": "Amazon Bedrock API key is expired",
|
||||
"CheckType": [
|
||||
"Software and Configuration Checks",
|
||||
"Industry and Regulatory Standards"
|
||||
"Software and Configuration Checks/AWS Security Best Practices",
|
||||
"Software and Configuration Checks/Industry and Regulatory Standards/AWS Foundational Security Best Practices",
|
||||
"Software and Configuration Checks/Industry and Regulatory Standards/CIS AWS Foundations Benchmark",
|
||||
"TTPs/Initial Access/Valid Accounts"
|
||||
],
|
||||
"ServiceName": "bedrock",
|
||||
"SubServiceName": "",
|
||||
"ResourceIdTemplate": "arn:partition:iam:region:account-id:user/{user-name}/credential/{api-key-id}",
|
||||
"ResourceIdTemplate": "",
|
||||
"Severity": "high",
|
||||
"ResourceType": "AwsIamServiceSpecificCredential",
|
||||
"ResourceType": "AwsIamUser",
|
||||
"ResourceGroup": "IAM",
|
||||
"Description": "Ensure that Amazon Bedrock API keys have expiration dates set to prevent long-term credential exposure. Long-term credentials pose a significant security risk as they remain valid indefinitely and can be used for unauthorized access if compromised.",
|
||||
"Risk": "Amazon Bedrock API keys without expiration dates are long-term credentials that remain valid indefinitely. This increases the risk of unauthorized access if the credentials are compromised, as they cannot be automatically invalidated. Long-term credentials violate the principle of credential rotation and can lead to security vulnerabilities, data breaches, or unauthorized usage of Bedrock services.",
|
||||
"RelatedUrl": "https://docs.aws.amazon.com/bedrock/latest/userguide/api-keys.html",
|
||||
"Description": "**Bedrock API keys** are evaluated for **lifetime** and **expiration**.\n\nThe finding identifies keys that are long-lived, set to expire far in the future, or configured to `never expire`, and distinguishes them from keys that have already expired.",
|
||||
"Risk": "Long-lived or non-expiring keys enable persistent access if compromised.\n- Confidentiality: unauthorized inference and exposure of prompts/outputs\n- Availability/Cost: uncontrolled usage and spend spikes\n- Integrity: actions can continue without timely revocation or rotation",
|
||||
"RelatedUrl": "",
|
||||
"AdditionalURLs": [
|
||||
"https://docs.aws.amazon.com/ja_jp/bedrock/latest/userguide/getting-started-api-keys.html",
|
||||
"https://docs.aws.amazon.com/IAM/latest/UserGuide/best-practices.html#rotate-credentials",
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/api-keys.html"
|
||||
],
|
||||
"Remediation": {
|
||||
"Code": {
|
||||
"CLI": "aws iam delete-service-specific-credential --user-name <username> --service-specific-credential-id <credential-id>",
|
||||
"NativeIaC": "",
|
||||
"Other": "",
|
||||
"Other": "1. Sign in to the AWS Management Console and open IAM\n2. Go to Users > select <example_resource_name> > Security credentials\n3. In \"API keys for Amazon Bedrock\", find the non-expired key and click Delete\n4. Confirm deletion to remove the key (removes the long-term credential so the check passes)",
|
||||
"Terraform": ""
|
||||
},
|
||||
"Recommendation": {
|
||||
"Text": "Delete the long-term API keys for Amazon Bedrock. Instead, use temporary credentials, IAM roles, or create new API keys with appropriate expiration dates. Implement a credential rotation policy to ensure all API keys have reasonable expiration periods. Consider using AWS STS for temporary credentials when possible.",
|
||||
"Url": "https://docs.aws.amazon.com/IAM/latest/UserGuide/best-practices.html#rotate-credentials"
|
||||
"Text": "Prefer **short-term credentials** and **IAM roles**; avoid `never expire`.\n\nEnforce **least privilege**, strict **rotation**, and automatic **expiration** for any long-term key. Store secrets securely, monitor with audit logs, and revoke unused or stale keys quickly.",
|
||||
"Url": "https://hub.prowler.com/check/bedrock_api_key_no_long_term_credentials"
|
||||
}
|
||||
},
|
||||
"Categories": [
|
||||
"gen-ai",
|
||||
"trust-boundaries"
|
||||
"secrets",
|
||||
"gen-ai"
|
||||
],
|
||||
"DependsOn": [],
|
||||
"RelatedTo": [],
|
||||
|
||||
+22
-12
@@ -1,27 +1,37 @@
|
||||
{
|
||||
"Provider": "aws",
|
||||
"CheckID": "bedrock_guardrail_prompt_attack_filter_enabled",
|
||||
"CheckTitle": "Configure Prompt Attack Filter with the highest strength for Amazon Bedrock Guardrails.",
|
||||
"CheckType": [],
|
||||
"CheckTitle": "Amazon Bedrock guardrail has prompt attack filter strength set to HIGH",
|
||||
"CheckType": [
|
||||
"Software and Configuration Checks/AWS Security Best Practices/Runtime Behavior Analysis",
|
||||
"TTPs/Defense Evasion",
|
||||
"Effects/Data Exposure"
|
||||
],
|
||||
"ServiceName": "bedrock",
|
||||
"SubServiceName": "",
|
||||
"ResourceIdTemplate": "arn:partition:bedrock:region:account-id:guardrails/resource-id",
|
||||
"ResourceIdTemplate": "",
|
||||
"Severity": "high",
|
||||
"ResourceType": "Other",
|
||||
"ResourceGroup": "ai_ml",
|
||||
"Description": "Ensure that prompt attack filter strength is set to HIGH for Amazon Bedrock guardrails to mitigate prompt injection and bypass techniques.",
|
||||
"Risk": "If prompt attack filter strength is not set to HIGH, Bedrock models may be more vulnerable to prompt injection attacks or jailbreak attempts, which could allow harmful or sensitive content to bypass filters and reach end users.",
|
||||
"RelatedUrl": "https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails.html",
|
||||
"Description": "**Bedrock guardrails** have the **Prompt attack** filter set to `HIGH` strength to detect and block injection and jailbreak patterns. Guardrails missing this setting or using lower strengths are identified.",
|
||||
"Risk": "Without **HIGH** prompt-attack filtering, models are exposed to **prompt injection/jailbreaks**:\n- Confidentiality: coerced disclosure of sensitive data\n- Integrity: policy evasion and manipulated outputs\n- Operations: unintended tool execution and workflow tampering",
|
||||
"RelatedUrl": "",
|
||||
"AdditionalURLs": [
|
||||
"https://trendmicro.com/cloudoneconformity/knowledge-base/aws/Bedrock/prompt-attack-strength.html",
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/prompt-injection.html",
|
||||
"https://support.icompaas.com/support/solutions/articles/62000233535-ensure-prompt-attack-filter-is-configured-at-highest-strength-for-amazon-bedrock-guardrails",
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails.html"
|
||||
],
|
||||
"Remediation": {
|
||||
"Code": {
|
||||
"CLI": "aws bedrock put-guardrails-configuration --guardrails-config 'promptAttackStrength=HIGH'",
|
||||
"NativeIaC": "",
|
||||
"Other": "https://www.trendmicro.com/cloudoneconformity/knowledge-base/aws/Bedrock/prompt-attack-strength.html",
|
||||
"Terraform": ""
|
||||
"CLI": "aws bedrock update-guardrail --guardrail-identifier <guardrail_id> --content-policy-config 'filtersConfig=[{type=PROMPT_ATTACK,inputStrength=HIGH}]'",
|
||||
"NativeIaC": "```yaml\nResources:\n <example_resource_name>:\n Type: AWS::Bedrock::Guardrail\n Properties:\n Name: <example_resource_name>\n BlockedInputMessaging: \"Blocked\"\n BlockedOutputsMessaging: \"Blocked\"\n ContentPolicyConfig:\n FiltersConfig:\n - Type: PROMPT_ATTACK # Critical: enables the Prompt Attack filter\n InputStrength: HIGH # Critical: sets filter strength to HIGH to pass the check\n```",
|
||||
"Other": "1. Open the AWS Console and go to Amazon Bedrock\n2. Select Guardrails, then choose your guardrail\n3. In Content filters, find Prompt attacks\n4. Set Strength to High\n5. Click Save",
|
||||
"Terraform": "```hcl\nresource \"aws_bedrock_guardrail\" \"<example_resource_name>\" {\n name = \"<example_resource_name>\"\n blocked_input_messaging = \"Blocked\"\n blocked_outputs_messaging = \"Blocked\"\n\n content_policy_config {\n filters_config {\n type = \"PROMPT_ATTACK\" # Critical: enables the Prompt Attack filter\n input_strength = \"HIGH\" # Critical: sets filter strength to HIGH to pass the check\n }\n }\n}\n```"
|
||||
},
|
||||
"Recommendation": {
|
||||
"Text": "Set the prompt attack filter strength to HIGH for Amazon Bedrock guardrails to prevent prompt injection attacks and ensure robust protection against content manipulation.",
|
||||
"Url": "https://docs.aws.amazon.com/bedrock/latest/userguide/prompt-injection.html"
|
||||
"Text": "Set the **Prompt attack** filter to `HIGH` and apply **defense in depth**:\n- Tag user/external inputs as untrusted for evaluation\n- Combine with denied topics and sensitive-info filters\n- Enforce **least privilege** and approvals for risky actions\n- Monitor guardrail hits and tune to reduce false negatives",
|
||||
"Url": "https://hub.prowler.com/check/bedrock_guardrail_prompt_attack_filter_enabled"
|
||||
}
|
||||
},
|
||||
"Categories": [
|
||||
|
||||
+20
-10
@@ -1,27 +1,37 @@
|
||||
{
|
||||
"Provider": "aws",
|
||||
"CheckID": "bedrock_guardrail_sensitive_information_filter_enabled",
|
||||
"CheckTitle": "Configure Sensitive Information Filters for Amazon Bedrock Guardrails.",
|
||||
"CheckType": [],
|
||||
"CheckTitle": "Amazon Bedrock guardrail blocks or masks sensitive information",
|
||||
"CheckType": [
|
||||
"Software and Configuration Checks/AWS Security Best Practices",
|
||||
"Software and Configuration Checks/AWS Security Best Practices/Runtime Behavior Analysis",
|
||||
"Effects/Data Exposure",
|
||||
"Sensitive Data Identifications/PII"
|
||||
],
|
||||
"ServiceName": "bedrock",
|
||||
"SubServiceName": "",
|
||||
"ResourceIdTemplate": "arn:partition:bedrock:region:account-id:guardrails/resource-id",
|
||||
"ResourceIdTemplate": "",
|
||||
"Severity": "high",
|
||||
"ResourceType": "Other",
|
||||
"ResourceGroup": "ai_ml",
|
||||
"Description": "Ensure that sensitive information filters are enabled for Amazon Bedrock guardrails to prevent the leakage of sensitive data such as personally identifiable information (PII), financial data, or confidential corporate information.",
|
||||
"Risk": "If sensitive information filters are not enabled, Bedrock models may inadvertently generate or expose confidential or sensitive information in responses, leading to data breaches, regulatory violations, or reputational damage.",
|
||||
"RelatedUrl": "https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails.html",
|
||||
"Description": "**Bedrock guardrails** use **sensitive information filters** to `block` or `mask` detected PII and custom pattern matches in prompts and responses.\n\nThe evaluation looks for guardrails with this filtering configured.",
|
||||
"Risk": "Absent filtering, prompts or outputs can reveal **PII**, credentials, or financial records, compromising **confidentiality**.\n- Exposed tokens enable unauthorized access and data tampering (integrity)\n- Disclosed customer details facilitate fraud and identity theft, with potential lateral movement",
|
||||
"RelatedUrl": "",
|
||||
"AdditionalURLs": [
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails-sensitive-filters.html",
|
||||
"https://www.trendmicro.com/cloudoneconformity/knowledge-base/aws/Bedrock/guardrails-with-pii-mask-block.html",
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails.html"
|
||||
],
|
||||
"Remediation": {
|
||||
"Code": {
|
||||
"CLI": "aws bedrock put-guardrails-configuration --guardrails-config 'sensitiveInformationFilter=true'",
|
||||
"CLI": "aws bedrock update-guardrail --guardrail-identifier <example_resource_id> --sensitive-information-policy-config '{\"piiEntitiesConfig\":[{\"type\":\"EMAIL\",\"action\":\"ANONYMIZE\"}]}'",
|
||||
"NativeIaC": "",
|
||||
"Other": "https://www.trendmicro.com/cloudoneconformity/knowledge-base/aws/Bedrock/guardrails-with-pii-mask-block.html",
|
||||
"Other": "1. Sign in to the AWS Console and open Amazon Bedrock\n2. Go to Guardrails and select <example_resource_name>\n3. Click Edit (or Open draft) and open Sensitive information filters\n4. Add PII type EMAIL and set action to Mask (or Block)\n5. Click Save",
|
||||
"Terraform": ""
|
||||
},
|
||||
"Recommendation": {
|
||||
"Text": "Enable sensitive information filters for Amazon Bedrock guardrails to prevent the exposure of sensitive or confidential information.",
|
||||
"Url": "https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails-sensitive-filters.html"
|
||||
"Text": "Enable and tune **sensitive information filters** for inputs and outputs.\n- Use `BLOCK` for high-risk disclosures; `ANONYMIZE` when context is needed\n- Add custom regex for org-specific IDs\n- Apply least privilege and data minimization\n- Test regularly and monitor outcomes as part of defense-in-depth",
|
||||
"Url": "https://hub.prowler.com/check/bedrock_guardrail_sensitive_information_filter_enabled"
|
||||
}
|
||||
},
|
||||
"Categories": [
|
||||
|
||||
+18
-10
@@ -1,27 +1,35 @@
|
||||
{
|
||||
"Provider": "aws",
|
||||
"CheckID": "bedrock_model_invocation_logging_enabled",
|
||||
"CheckTitle": "Ensure that model invocation logging is enabled for Amazon Bedrock.",
|
||||
"CheckType": [],
|
||||
"CheckTitle": "Amazon Bedrock model invocation logging is enabled",
|
||||
"CheckType": [
|
||||
"Software and Configuration Checks/AWS Security Best Practices",
|
||||
"Software and Configuration Checks/Industry and Regulatory Standards/AWS Foundational Security Best Practices"
|
||||
],
|
||||
"ServiceName": "bedrock",
|
||||
"SubServiceName": "",
|
||||
"ResourceIdTemplate": "arn:partition:bedrock:region:account-id:model/resource-id",
|
||||
"ResourceIdTemplate": "",
|
||||
"Severity": "medium",
|
||||
"ResourceType": "Other",
|
||||
"ResourceGroup": "ai_ml",
|
||||
"Description": "Ensure that model invocation logging is enabled for Amazon Bedrock service in order to collect metadata, requests, and responses for all model invocations in your AWS cloud account.",
|
||||
"Risk": "In Amazon Bedrock, model invocation logging enables you to collect the invocation request and response data, along with metadata, for all 'Converse', 'ConverseStream', 'InvokeModel', and 'InvokeModelWithResponseStream' API calls in your AWS account. Each log entry includes important details such as the timestamp, request ID, model ID, and token usage. Invocation logs can be utilized for troubleshooting, performance enhancements, abuse detection, and security auditing. By default, model invocation logging is disabled.",
|
||||
"RelatedUrl": "https://docs.aws.amazon.com/bedrock/latest/userguide/model-invocation-logging.html",
|
||||
"Description": "**Bedrock** model invocation logging captures request, response, and metadata for `Converse`, `ConverseStream`, `InvokeModel`, and `InvokeModelWithResponseStream` calls per Region, delivering records to **CloudWatch Logs** and/or **S3** when configured.",
|
||||
"Risk": "Without **invocation logs**, you lose **auditability** and **forensic visibility** into model activity.\n\nCredential misuse or **prompt injection/jailbreak** attempts may go unnoticed, enabling data exfiltration and unauthorized spend. Missing traceability weakens **integrity** controls and slows incident response.",
|
||||
"RelatedUrl": "",
|
||||
"AdditionalURLs": [
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/model-invocation-logging.html#model-invocation-logging-console",
|
||||
"https://www.trendmicro.com/cloudoneconformity/knowledge-base/aws/Bedrock/enable-model-invocation-logging.html",
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/model-invocation-logging.html"
|
||||
],
|
||||
"Remediation": {
|
||||
"Code": {
|
||||
"CLI": "aws bedrock put-model-invocation-logging-configuration --logging-config 's3Config={bucketName='tm-bedrock-logging-data',keyPrefix='invocation-logs'},textDataDeliveryEnabled=true,imageDataDeliveryEnabled=true,embeddingDataDeliveryEnabled=true'",
|
||||
"CLI": "aws bedrock put-model-invocation-logging-configuration --logging-config '{\"s3Config\":{\"bucketName\":\"<example_resource_name>\"},\"textDataDeliveryEnabled\":true}'",
|
||||
"NativeIaC": "",
|
||||
"Other": "https://www.trendmicro.com/cloudoneconformity/knowledge-base/aws/Bedrock/enable-model-invocation-logging.html",
|
||||
"Other": "1. Open the Amazon Bedrock console in the target Region\n2. Go to Settings > Model invocation logging\n3. Toggle Logging to On\n4. Select Amazon S3 as the destination and choose <example_resource_name> bucket\n5. Under Data types, select Text\n6. Click Save",
|
||||
"Terraform": ""
|
||||
},
|
||||
"Recommendation": {
|
||||
"Text": "Enable model invocation logging for Amazon Bedrock service in order to collect metadata, requests, and responses for all model invocations in your AWS cloud account.",
|
||||
"Url": "https://docs.aws.amazon.com/bedrock/latest/userguide/model-invocation-logging.html#model-invocation-logging-console"
|
||||
"Text": "Enable **model invocation logging** and route events to **CloudWatch Logs** and/or **S3**.\n\nEnforce **least privilege** on log access, use encryption, and set retention/lifecycle policies. Monitor for anomalies and alerts to support **defense in depth** and **separation of duties**.",
|
||||
"Url": "https://hub.prowler.com/check/bedrock_model_invocation_logging_enabled"
|
||||
}
|
||||
},
|
||||
"Categories": [
|
||||
|
||||
+23
-12
@@ -1,27 +1,38 @@
|
||||
{
|
||||
"Provider": "aws",
|
||||
"CheckID": "bedrock_model_invocation_logs_encryption_enabled",
|
||||
"CheckTitle": "Ensure that Amazon Bedrock model invocation logs are encrypted with KMS.",
|
||||
"CheckType": [],
|
||||
"CheckTitle": "Amazon Bedrock model invocation logs are encrypted in the S3 bucket and KMS-encrypted in the CloudWatch log group",
|
||||
"CheckType": [
|
||||
"Software and Configuration Checks/AWS Security Best Practices",
|
||||
"Software and Configuration Checks/Industry and Regulatory Standards/AWS Foundational Security Best Practices",
|
||||
"Software and Configuration Checks/Industry and Regulatory Standards/CIS AWS Foundations Benchmark",
|
||||
"Effects/Data Exposure"
|
||||
],
|
||||
"ServiceName": "bedrock",
|
||||
"SubServiceName": "",
|
||||
"ResourceIdTemplate": "arn:partition:bedrock:region:account-id:model/resource-id",
|
||||
"ResourceIdTemplate": "",
|
||||
"Severity": "high",
|
||||
"ResourceType": "Other",
|
||||
"ResourceType": "AwsS3Bucket",
|
||||
"ResourceGroup": "ai_ml",
|
||||
"Description": "Ensure that Amazon Bedrock model invocation logs are encrypted using AWS KMS to protect sensitive data in the request and response logs for all model invocations.",
|
||||
"Risk": "If Amazon Bedrock model invocation logs are not encrypted, sensitive data such as prompts, responses, and token usage could be exposed to unauthorized parties. This may lead to data breaches, security vulnerabilities, or unintended use of sensitive information.",
|
||||
"RelatedUrl": "https://docs.aws.amazon.com/bedrock/latest/userguide/data-protection.html",
|
||||
"Description": "**Bedrock model invocation logs** are stored in encrypted destinations: **S3 buckets** with bucket encryption and **CloudWatch Logs** groups protected by an AWS KMS key.\n\nThis evaluates whether configured log targets enforce encryption at rest for request/response content and associated metadata.",
|
||||
"Risk": "Without encryption at rest, prompts, outputs, images, and token/usage metadata in logs can be read if S3 or CloudWatch storage or replicas are accessed by unauthorized principals.\n\nImpacts:\n- Loss of data **confidentiality**\n- Secret or PII exposure enabling **account abuse**\n- Operational intel for **lateral movement**",
|
||||
"RelatedUrl": "",
|
||||
"AdditionalURLs": [
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/data-protection.html",
|
||||
"https://docs.aws.amazon.com/bedrock/latest/userguide/model-invocation-logging.html",
|
||||
"https://support.icompaas.com/support/solutions/articles/62000233532-ensure-that-amazon-bedrock-model-invocation-logs-are-encrypted-with-kms",
|
||||
"https://docs.aws.amazon.com/prescriptive-guidance/latest/patterns/configure-bedrock-invocation-logging-cloudformation.html"
|
||||
],
|
||||
"Remediation": {
|
||||
"Code": {
|
||||
"CLI": "",
|
||||
"NativeIaC": "",
|
||||
"Other": "",
|
||||
"Terraform": ""
|
||||
"NativeIaC": "```yaml\nResources:\n EncryptedLogsBucket:\n Type: AWS::S3::Bucket\n Properties:\n BucketName: <example_resource_name>\n BucketEncryption: # critical: enables default encryption for the bucket (SSE-S3)\n ServerSideEncryptionConfiguration:\n - ServerSideEncryptionByDefault:\n SSEAlgorithm: AES256\n\n EncryptedLogGroup:\n Type: AWS::Logs::LogGroup\n Properties:\n LogGroupName: <example_resource_name>\n KmsKeyId: <example_kms_key_arn> # critical: encrypts the CloudWatch log group with a KMS key\n```",
|
||||
"Other": "1. In the Bedrock console, go to Settings and note the S3 bucket and CloudWatch log group used for Model invocation logging.\n2. S3 bucket: AWS Console > S3 > Buckets > <bucket> > Properties > Default encryption > Enable > Choose SSE-S3 (AES-256) > Save.\n3. CloudWatch Logs: AWS Console > CloudWatch > Logs > Log groups > select <log group> > Actions > Edit > KMS encryption > select <KMS key> > Save.",
|
||||
"Terraform": "```hcl\nresource \"aws_s3_bucket\" \"example\" {\n bucket = \"<example_resource_name>\"\n}\n\nresource \"aws_s3_bucket_server_side_encryption_configuration\" \"example\" {\n bucket = aws_s3_bucket.example.id\n rule {\n apply_server_side_encryption_by_default {\n sse_algorithm = \"AES256\" # critical: enables default encryption for the S3 bucket\n }\n }\n}\n\nresource \"aws_cloudwatch_log_group\" \"example\" {\n name = \"<example_resource_name>\"\n kms_key_id = \"<example_kms_key_arn>\" # critical: encrypts the log group with a KMS key\n}\n```"
|
||||
},
|
||||
"Recommendation": {
|
||||
"Text": "Ensure that model invocation logs for Amazon Bedrock are encrypted using AWS KMS to prevent unauthorized access to sensitive log data.",
|
||||
"Url": "hhttps://docs.aws.amazon.com/bedrock/latest/userguide/data-protection.html"
|
||||
"Text": "Ensure all invocation logs are encrypted end to end:\n- Enable S3 default encryption, preferably `SSE-KMS`, and restrict key usage\n- Assign a KMS key to CloudWatch log groups\n- Enforce **least privilege** on keys and logs, rotate keys, and monitor access for **defense in depth**",
|
||||
"Url": "https://hub.prowler.com/check/bedrock_model_invocation_logs_encryption_enabled"
|
||||
}
|
||||
},
|
||||
"Categories": [
|
||||
|
||||
Reference in New Issue
Block a user