create-dead-letter-queue
Pass
Audited by Gen Agent Trust Hub on Mar 17, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No evidence of instructions attempting to bypass safety filters or override agent behavior. The language is purely instructional and focused on code generation.
- [DATA_EXFILTRATION]: No hardcoded credentials or sensitive file path access detected. The provided
.envexamples use safe placeholders for configuration. - [EXTERNAL_DOWNLOADS]: The skill references well-known and trusted PHP libraries (e.g., Ramsey/Uuid, Symfony/Console, PhpAmqpLib) which are standard for the described tasks.
- [COMMAND_EXECUTION]: No dangerous system commands or arbitrary code execution patterns were found. The generated console commands are for internal application management.
- [OBFUSCATION]: The code is provided in plain text with no evidence of encoding, zero-width characters, or hidden payloads.
- [INDIRECT_PROMPT_INJECTION]: While the skill generates code that processes external message data (message bodies from a queue), it follows standard development patterns (JSON decoding) and does not interpolate this data into agent prompts or shell commands in a way that suggests a vulnerability to the agent itself.
Audit Metadata