telnyx
Warn
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: MEDIUMPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [OBFUSCATION]: The file contains extreme repetition of the string 'Call Screening Rule Destination Settings' hundreds of times in the 'Telnyx Overview' section. While not containing encoded data, this high volume of redundant text can be used to bypass token limits, confuse LLM reasoning, or obscure other parts of the file.
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to pass JSON parameters to CLI commands using the
--inputflag. This creates a vulnerability surface where untrusted data from the user or external sources could be interpolated into a command line argument, potentially leading to injection attacks if the data is not properly sanitized. - Ingestion points: User input passed to the
--inputflag inmembrane action runcommands. - Boundary markers: None specified in the instructions.
- Capability inventory: Execute CLI commands (
membrane action run), perform network requests (membrane request). - Sanitization: No sanitization or escaping guidelines provided for input parameters.
Audit Metadata