The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions override the AI agent's typical conversational model by enforcing an autonomous execution mode that significantly reduces human oversight during destructive operations.
The instructions explicitly command: "Phases 1–6 run fully autonomously — do NOT ask the user any questions."
It further dictates that the agent should only present issues to the user if a failure persists after three autonomous fix attempts, effectively concealing interim code changes and errors.
[COMMAND_EXECUTION]: The skill orchestrates complex filesystem and network operations through a series of shell and Python scripts that modify the local environment.
Phase 4 involves autonomous, file-by-file modifications to the project's source code to replace Twilio logic with Telnyx implementations.
The skill uses various scripts (e.g., run-discovery.sh, validate-migration.sh) to perform deep scans and transformations.
The script scripts/test-migration/webhook-receiver.py utilizes subprocess.Popen to manage local test processes.
[EXTERNAL_DOWNLOADS]: The skill automates the installation of several developer tools and libraries.
It initiates pip install, npm install, and go get for Telnyx-branded SDKs across multiple languages.
While these packages originate from the official vendor (team-telnyx), the autonomous nature of the skill means these installations occur without individual confirmation prompts.

telnyx-twilio-migration