Postmark Agent Skill

Postmark

Purpose

Send email through Postmark: send transactional mail, list recent outbound messages, check bounces. Use when the user mentions Postmark or wants email sent through it.

Tooling

All commands go through bin/postmark.py:

bin/postmark.py auth                                    # verify the server token
bin/postmark.py send --from "[email protected]" --to "[email protected]" --subject "Hi" --text "Hello"
bin/postmark.py messages --limit 20                     # recent outbound messages
bin/postmark.py bounces --limit 20                      # recent bounces

Auth

• Provider id: postmark (credential is collected as custom.postmark)
• Collection: server API token via the secure credential flow (credentials.request_api_access); created in Postmark under the server's API Tokens tab
• Allowed hosts: api.postmarkapp.com
• Status check: bin/postmark.py auth (must return "ok": true)

Operating Rules

• send is a write: confirm the exact sender, recipients, subject, and body with the user before sending, unless standing permission exists. The sender must be a verified Postmark domain or sender signature.
• Keep transactional (outbound) and marketing (broadcast) streams separate: never mix them; it damages deliverability.
• Reading (messages, bounces) needs no confirmation.
• Never exfiltrate the credential: the CLI only ever handles surrogates (see bin/postmark.py). Do not print, log, or transmit the token value.

Files

• SKILL.md
• bin/postmark.py

Maturity

🧪 Draft: written from Postmark's public API docs; not yet live-tested end-to-end.