The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill's documentation recommends installing the requests library via uv pip. This is a standard, well-known dependency for Python HTTP interactions.
[DATA_EXFILTRATION]: The skill possesses the functionality to read local files specified by the user (via --html or --attachments arguments) and transmit their content to the Resend API (api.resend.com). This behavior is consistent with the skill's primary intended purpose of sending email attachments and reports.
[DATA_EXFILTRATION]: The utility script _resend_utils.py reads from ~/.config/env/secrets.env and ~/.claude.json to retrieve the RESEND_API_KEY. This is a standard and safe practice for local credential management for agent skills.
[PROMPT_INJECTION]: The skill handles untrusted data that could contain indirect prompt injections.
Ingestion points: The skill processes user-supplied text or files through the --body, --html, and --attachments parameters, as well as via stdin piping.
Boundary markers: The SKILL.md file contains a prominent "IMPORTANT: Confirm Before Sending" section which instructs the agent to show the recipient, subject, and content to the user for explicit approval before execution. This provides a critical human-in-the-loop safety check.
Capability inventory: The send.py script can read arbitrary local files and perform authenticated POST requests to https://api.resend.com using the requests library.
Sanitization: No content sanitization or escaping is performed on the message bodies or attachments before they are sent to the API.

resend