The Agent Skills Directory

[COMMAND_EXECUTION]: The skill employs the swan-execute-code tool to process attendee data in batches and generate sequence actions. This is used for firmographic filtering (Step 3) and bulk personalization (Step 8).
[EXTERNAL_DOWNLOADS]: The skill retrieves data from external sources, including event platforms (via webhooks) and public event websites (via Apify scraping). Apify is a well-known service used to gather attendee information from industry and competitor events.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its ingestion of untrusted content from external websites and attendee lists. This data is interpolated into prompts for lead qualification and email drafting.
Ingestion points: Webhooks from event platforms, CSV uploads, and web scrapes performed via Apify.
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within external content.
Capability inventory: The agent has access to swan-execute-code, swan-update-skill, CRM integrations (HubSpot), and automated messaging tools (reach-out, swan-build-sequence).
Sanitization: No explicit sanitization or validation of the ingested external content is mentioned.
[COMMAND_EXECUTION]: The skill demonstrates self-modification capabilities by using swan-update-skill to rewrite its own SKILL.md file. This is utilized to persist configuration details, such as trigger sources and ICP filters, in the 'Setup state' section.

event-radar