The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill outlines a system for ingesting and processing data from untrusted external sources, creating a potential attack surface for indirect prompt injection.
Ingestion points: The DataSourceManager and specialized functions like scrape_linkedin_profile, scrape_job_postings, and get_linkedin_activity are designed to pull data from external platforms (LinkedIn, news sites, APIs) (SKILL.md).
Boundary markers: The instructions and provided Python templates do not include boundary markers or explicit instructions for the LLM to ignore potentially malicious embedded commands within the fetched prospect data.
Capability inventory: While the skill primarily focuses on data synthesis and scoring, the process_event logic includes an alert mechanism (send_trigger_alert) and recommendation logic (get_recommended_action) that could be influenced by injected content.
Sanitization: There is no evidence of sanitization or filtering logic for the raw text ingested from external sources before it is analyzed for insights (e.g., in extract_topics or analyze_content_style).

prospect-research-integration