The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill documents the installation of the Tailscale client using an official shell script in AGENT-INBOX.md: curl -fsSL https://tailscale.com/install.sh | sh. This script is used to set up secure tunnels for testing webhook endpoints locally.\n- [PRIVILEGE_ESCALATION]: Instructions for configuring development tunnels in AGENT-INBOX.md include commands that require administrative privileges using sudo (e.g., sudo tailscale up, sudo tailscale funnel 3000). While typical for managing system-level networking tools, these commands grant elevated access to the environment.\n- [EXTERNAL_DOWNLOADS]: The skill provides instructions to download and install the Resend CLI and various SDKs from public package registries. It also mentions third-party development tools such as ngrok, cloudflared, and localtunnel.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process incoming emails, which constitutes a surface for indirect prompt injection.\n
Ingestion points: Incoming email data is received via inbox.email.received and email.received webhooks, as described in AGENT-INBOX.md and RECEIVING.md.\n
Boundary markers: The skill instructs the agent to treat received content as untrusted data and provides a security framework with multiple levels of filtering and allowlists.\n
Capability inventory: The Resend SDK enables the agent to read and respond to emails, manage contact segments, and configure topics.\n
Sanitization: The skill recommends stripping quoted content, using text-only processing to remove HTML, and applying safety patterns based on OWASP guidelines.

resend