The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to fetch and analyze external web content, which serves as a vector for indirect prompt injection.
Ingestion points: The scripts/fetch_page.py utility is used to download content from arbitrary public URLs provided by the user or discovered during search analysis.
Boundary markers: The skill contains explicit instructions in SKILL.md and references/skill-contract.md to treat scraped material as untrusted input and to ignore any instructions embedded within that data.
Capability inventory: The skill possesses network read capabilities (scripts/fetch_page.py) and file-write capabilities for generating reports and schemas (scripts/generate_markdown_report.py, scripts/generate_pdf_report.py, and scripts/schema_generator.py).
Sanitization: The fetch_page.py script performs basic HTML metadata extraction and unescaping but does not execute or evaluate the content of the pages it fetches.
[DATA_EXFILTRATION]: The skill includes a utility scripts/fetch_page.py that performs network operations (HTTP/HTTPS GET requests) to arbitrary public domains.
Evidence: The script uses urllib.request to fetch page content. While it includes robust SSRF protection by validating that hostnames do not resolve to private, loopback, or reserved IP addresses, it still allows communication with any external public server.

serp-analysis