The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions provide commands to install the Porteden CLI from an external GitHub repository (github.com/porteden/cli/cmd/porteden@latest) via go install and a Homebrew tap (porteden/tap/porteden).
[COMMAND_EXECUTION]: The skill uses the porteden binary to perform various email operations including listing, reading, and sending messages.
[DATA_EXPOSURE]: The skill accesses and manages sensitive email data across multiple providers. It utilizes an environment variable PE_API_KEY for authentication and stores credentials in the system keyring.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from email messages and threads, creating a surface for indirect prompt injection.
Ingestion points: Email content is ingested into the agent context through the porteden email message and porteden email thread commands as specified in SKILL.md.
Boundary markers: The skill does not implement boundary markers or instructions to the agent to ignore embedded commands within the email data.
Capability inventory: The agent possesses extensive capabilities including send, reply, forward, modify, and delete, which could be abused if malicious instructions are present in processed emails.
Sanitization: There is no evidence of sanitization, filtering, or validation of the email content before it is processed by the agent.

porteden-email