The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches an installation script from Tailscale's official website to set up networking capabilities.
[REMOTE_CODE_EXECUTION]: Executes the downloaded Tailscale installation script directly through the shell.
[PRIVILEGE_ESCALATION]: Employs sudo to perform system-level tasks, including installing Tailscale and configuring the network interface, which are necessary for the skill's intended functionality.
[COMMAND_EXECUTION]: Utilizes system commands such as tailscale status, tailscale ip -4, and tailscale up to manage the local environment's network status.
[INDIRECT_PROMPT_INJECTION]: The skill is primarily designed to ingest and act upon data from external webhooks, establishing an attack surface for indirect injection.
Ingestion points: External POST requests sent to a webhook task, as described in SKILL.md.
Boundary markers: The instructions recommend using specific placeholders like {{body.action}} to limit the scope of external data exposure.
Capability inventory: The skill can schedule tasks (schedule_task), manage secrets (request_secret), drive browser interactions (preview_open), and execute shell commands for networking.
Sanitization: Explicitly instructs the agent to treat all rendered webhook values as untrusted data and strictly avoid interpreting them as instructions.

make-bot-ui