The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill fetches historical page titles and meta descriptions from the Wayback Machine (Step 4b) and interpolates this untrusted external data directly into a prompt for the Gemini LLM (Step 4e) without specific sanitization or boundary markers. This creates an attack surface where malicious content embedded in archived websites could influence the agent's behavior.
[COMMAND_EXECUTION]: The skill uses multiple inline Python scripts (python3 -c) to process data and normalize inputs. While the scripts themselves are static and provided within the instructions, they interpolate user-supplied domain lists and external API responses, which requires the agent to handle escaping correctly to avoid injection into the shell environment.
[EXTERNAL_DOWNLOADS]: The skill fetches data from well-known and reputable services, including the Wayback Machine (archive.org), RDAP.org for domain registration lookups, and the Google Gemini API (googleapis.com) for relevance analysis. These operations are transparently documented and follow rate-limiting best practices.
[SAFE]: The skill implements a robust 'Conservative-by-default' rule and an anti-abuse policy that discourages the creation of Private Blog Networks (PBNs) and deceptive redirects. It also enforces a mandatory disclaimer on all outputs, ensuring users are aware of the risks involved in expired domain acquisition.

domain-expired-opportunity-finder