[ 
https://issues.apache.org/jira/browse/CAMEL-23250?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18068563#comment-18068563
 ] 

Guillaume Nodet commented on CAMEL-23250:
-----------------------------------------

**Security Audit Findings — Additional Recommendations**

After auditing the codebase for CAMEL-22497 (easy HTTPS), here are additional 
areas that should be addressed as part of this ticket:

*High Severity — Dev-only settings that disable security checks:*

|| Setting || Scope || Notes ||
| {{trustAllCertificates}} | 37+ AWS components | Disables all SSL certificate 
validation, no runtime warning |
| {{sslValidationEnabled=false}} | MongoDB ({{SslAwareMongoClient}}) | 
Hardcoded TrustManager that accepts everything + disables hostname verification 
|
| {{selfSigned=true}} | Core SSL ({{SSLConfigurationProperties}}) | Already 
warns at startup (added in CAMEL-22497) |

*Medium Severity — Insecure defaults or bypassable checks:*

|| Setting || Scope || Notes ||
| {{tls=false}} (default) | Qdrant, other vector DB components | TLS disabled 
by default |
| {{httpsHostnameVerificationEnabled}} | PAHO MQTT / MQTT5 | Can be set to 
false to skip hostname verification |

*Recommendations:*

# Emit a prominent WARN log at startup whenever any of the above insecure 
settings are active (similar to what we already do for {{selfSigned=true}})
# Ensure all password/token/key fields across components have 
{{@Metadata(secret=true)}}. Most are correct (Kafka, AI components), but a 
systematic audit would catch gaps.
# Consider a "strict mode" or production profile that refuses insecure settings 
unless explicitly acknowledged
# All "disable security" properties should follow the same opt-in + warning 
pattern

_Claude Code on behalf of Guillaume Nodet_

> Warn or prevent plain-text secrets in configuration properties
> --------------------------------------------------------------
>
>                 Key: CAMEL-23250
>                 URL: https://issues.apache.org/jira/browse/CAMEL-23250
>             Project: Camel
>          Issue Type: Improvement
>          Components: camel-core
>            Reporter: Guillaume Nodet
>            Assignee: Guillaume Nodet
>            Priority: Major
>
> Currently, Camel has a well-established pipeline for detecting and masking 
> secret properties (via @Metadata(secret=true) and SensitiveUtils), but it 
> only masks values in logs and console output. There is no mechanism to warn 
> users when secrets are configured via plain-text properties instead of using 
> secure alternatives like RAW(), vault references ({{vault:...}}), or 
> environment variable placeholders (${env:...}).
> This improvement should:
> h3. 1. Warn on plain-text secrets
> *Log a warning at startup* when a secret property is set via plain-text (not 
> RAW(), not {{vault:...}}, not ${env:...}). This could be added in 
> MainHelper.sensitiveAwareLogging() or BaseMainSupport auto-configuration.
> h3. 2. Add a configuration flag
> Add a *camel.main.warnOnPlainTextSecrets* flag (default: true) to control 
> whether warnings are emitted, allowing users to suppress them in development 
> environments.
> h3. 3. Fix PropertiesDevConsole JSON output
> PropertiesDevConsole currently does *not mask secret values in JSON mode* 
> (only masks in text mode), potentially exposing secrets via monitoring 
> endpoints.
> h3. 4. Strict mode
> Consider adding a *strict mode* (e.g. camel.main.forbidPlainTextSecrets) that 
> would fail startup if plain-text secrets are detected, for production 
> hardening.
> h3. 5. Warn on development-only settings in production
> Beyond secrets, some configuration options are inherently unsafe for 
> production use, such as:
> - *camel.ssl.selfSigned=true* — generates an ephemeral self-signed 
> certificate (added in CAMEL-22497)
> - *camel.ssl.trustAllCertificates=true* — disables certificate validation
> These are not secrets (they are boolean flags) so @Metadata(secret=true) is 
> not the right mechanism. Consider adding a new annotation attribute like 
> @Metadata(label = "development") or @Metadata(warnInProduction = true) that 
> would trigger a startup warning when these options are enabled in a 
> non-development profile. This would catch cases where development settings 
> are accidentally left in production configuration.
> Related: there are currently 143+ secret keys detected by SensitiveUtils. The 
> detection infrastructure is solid - it just needs to be leveraged for 
> prevention, not just masking.
> Key files:
> - core/camel-main/src/main/java/org/apache/camel/main/MainHelper.java 
> (sensitiveAwareLogging)
> - core/camel-util/src/main/java/org/apache/camel/util/SensitiveUtils.java
> - 
> core/camel-console/src/main/java/org/apache/camel/impl/console/PropertiesDevConsole.java
> - 
> core/camel-main/src/main/java/org/apache/camel/main/SSLConfigurationProperties.java
>  (selfSigned, trustAllCertificates)



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to