telnyx-twilio-migration
Pass
Audited by Gen Agent Trust Hub on Mar 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill autonomously executes a series of shell scripts included in its package (e.g.,
run-discovery.sh,validate-migration.sh, andlint-telnyx-correctness.sh) to perform project analysis, linting, and validation. These operations are conducted without user intervention following the initial Phase 0 setup. The scripts use standard utilities likegrepandfindto analyze the project directory. - [PROMPT_INJECTION]: The skill presents a surface for Indirect Prompt Injection. It reads and processes arbitrary text from the user's existing Twilio codebase to identify patterns and perform code transformations. If the codebase contains malicious instructions embedded in comments or strings, the agent could potentially be misled during the transformation process.
- Ingestion points: Project source code files are read extensively during the discovery (Phase 1) and migration (Phase 4) loops.
- Boundary markers: The instructions do not specify strict delimiters to distinguish legitimate source code from potential adversarial instructions hidden within the project data.
- Capability inventory: The agent has high-tier capabilities, including the ability to modify project source files, write new documentation, and execute shell scripts.
- Sanitization: No explicit sanitization or filtering of the codebase content is described before the AI processes it for transformation logic.
Audit Metadata