Skills
skills/zlstas/skills/web-scraping-python/Gen Agent Trust Hub

web-scraping-python

Pass

Audited by Gen Agent Trust Hub on Apr 8, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill naturally interacts with untrusted web content which presents a surface for indirect prompt injection if scraped data is processed by an LLM without sanitization. Ingestion points: Network requests in examples/after.md and scripts/new_scraper.py. Boundary markers: Absent. Capability inventory: Network access for scraping and file system writes for data storage. Sanitization: The skill uses structural parsing but does not define sanitization for downstream LLM use.
  • [COMMAND_EXECUTION]: The scripts/new_scraper.py utility generates executable Python scripts and utilizes chmod(0o755) to set execution permissions, which is a standard pattern for developer scaffolding tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 8, 2026, 10:37 PM