The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill fetches and processes content from external URLs provided by the user, which presents a surface for indirect instructions to influence agent behavior.
Ingestion points: scripts/analyze_page.py and scripts/competitor_scanner.py use urllib to fetch the full HTML content of user-specified websites for analysis.
Boundary markers: The skill does not implement explicit boundary markers or instructions for the agent to ignore potentially malicious content embedded within the analyzed pages.
Capability inventory: The skill can perform network operations, write markdown files, and generate PDF reports locally.
Sanitization: The utility scripts extract text and metadata from HTML but do not perform security-focused sanitization of the content before presenting it to the agent.
[COMMAND_EXECUTION]: The orchestrator invokes several local Python scripts to perform analysis and report generation. These scripts are purpose-built for the skill and perform benign marketing analysis tasks.
[EXTERNAL_DOWNLOADS]: The scripts/generate_pdf_report.py script requires the reportlab library, which is a well-known and standard Python package for PDF creation.
[DATA_EXFILTRATION]: The scripts analyze_page.py and competitor_scanner.py perform network requests to arbitrary external domains provided by the user. Notably, these scripts explicitly disable SSL certificate verification (ssl.CERT_NONE), which is a poor security practice that could allow man-in-the-middle attacks to modify the data being ingested for analysis.

market