The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches external packages from the npm registry.
npx -y pigeonscale@0.0.25 in SKILL.md automatically downloads and runs the CLI.
pnpm add @pigeonscale/openclaw openclaw in references/openclaw.md installs third-party libraries.
[REMOTE_CODE_EXECUTION]: The use of npx -y to execute a specific version of a third-party package from the npm registry constitutes a remote code execution vector, as the package content is downloaded and executed at runtime.
[COMMAND_EXECUTION]: The skill instructions require the agent to execute multiple shell commands for mailbox management, authenticated sessions, and email operations using the pigeonscale CLI.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process email data from external sources, creating a surface for indirect prompt injection attacks.
Ingestion points: Incoming email content retrieved via pigeonscale mail list and pigeonscale mail watch (in SKILL.md).
Boundary markers: The instructions suggest that the agent clearly delimit email content as quoted external content.
Capability inventory: The skill has the ability to send emails, create accounts, and execute shell commands through the CLI.
Sanitization: The documentation describes platform-level mechanisms for flagging and redacting potentially malicious content in emails.

instant-agent-email-accounts-with-pigeonscale