telnyx-twilio-migration
Warn
Audited by Gen Agent Trust Hub on May 13, 2026
Risk Level: MEDIUMPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains strong imperative instructions directing the agent to operate autonomously during six of the seven migration phases. It explicitly commands the agent to "not ask the user any questions" and make decisions "deterministically." This pattern is an attempt to remove user oversight and bypass "human-in-the-loop" safety protocols during sensitive operations like file modification and API resource creation.
- Evidence (SKILL.md): "Phases 1–6 run fully autonomously — do NOT ask the user any questions. Make all decisions deterministically... This is the last time you will ask the user for input."
- [COMMAND_EXECUTION]: The skill requires the agent to execute numerous bash scripts included in its package. These scripts perform project-wide scanning, modify source files, and run integration tests that incur charges.
- Evidence (SKILL.md): "bash {baseDir}/scripts/run-discovery.sh ", "bash {baseDir}/scripts/validate-migration.sh ", and multiple integration test scripts in Phase 5.
- [COMMAND_EXECUTION]: The migration process involves collecting and exporting a Telnyx API key as an environment variable. While functional for the tool, the instructions to proceed without further user confirmation after Phase 0 increase the risk that the key or sensitive project data could be misused or exposed during the autonomous execution of the included shell scripts.
Audit Metadata