The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx to download and execute the @useclawlink/cli package from the npm registry. This package is the official command-line interface provided by the vendor for the Postmark integration.
[COMMAND_EXECUTION]: The skill interacts with the Postmark API by executing shell commands using the npx @useclawlink/cli tool. These commands facilitate sending emails, checking stats, and managing templates.
[CREDENTIALS_UNSAFE]: The skill documentation identifies that authentication credentials for the ClawLink service are stored locally on the user's filesystem at ~/.clawlink/credentials.json following the browser login process.
[INDIRECT_PROMPT_INJECTION]: The skill provides the ability to ingest data from external sources, such as bounce messages and delivery statistics, which could potentially contain adversarial content designed to influence the agent's behavior.
Ingestion points: Data is retrieved via actions like postmark_get_bounces and postmark_get_delivery_stats defined in SKILL.md.
Boundary markers: There are no explicit instructions for the agent to treat retrieved email data as untrusted or to use specific delimiters.
Capability inventory: The agent has the ability to send emails (postmark_send_email), modify templates (postmark_edit_template), and create webhooks (postmark_create_webhook).
Sanitization: No automated sanitization or verification of the content retrieved from Postmark is mentioned.

postmark