The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external websites and search results to identify prospects and draft outreach messages, creating a surface for indirect prompt injection.
Ingestion points: Data enters the agent's context through get_serp_results and manual browsing of prospect websites during the qualification and contact discovery phases.
Boundary markers: Absent. The instructions do not specify the use of delimiters or 'ignore' instructions for the external content being processed.
Capability inventory: The skill can write prospect lists to .csv files and outreach drafts to .md files in the local filesystem. It also has the ability to send data to a remote service via the create_share_report tool.
Sanitization: Absent. There are no instructions for sanitizing or escaping the content retrieved from prospect sites before it is included in files or reports.
[DATA_EXPOSURE]: The skill uses a create_share_report tool to publish findings to a hosted service. It contains explicit guardrails requiring user consent before sharing sensitive data like Search Console or Google Analytics numbers.
[EXTERNAL_DOWNLOADS]: The skill provides instructions to connect to a remote MCP server at app.harborrank.com. This is a legitimate interaction with the vendor's infrastructure for the skill's primary functionality.

link-prospecting