web-scraping-python

Warn

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The script 'scripts/new_scraper.py' generates executable Python code using a string template and writes it to the local filesystem. The script accepts a 'scraper-name' argument from the command line and uses it directly to construct the output file path (e.g., 'Path(f"{scraper_name}.py")'). Due to a lack of sanitization for directory traversal sequences like '../', this could allow an attacker to write or overwrite files outside the intended working directory. Furthermore, the script uses 'chmod' to modify file permissions on the generated file, which presents a risk if misused.
  • [PROMPT_INJECTION]: The skill is designed to ingest and process data from untrusted web sources, creating a surface for indirect prompt injection (Category 8). Evidence Chain: 1. Ingestion points: 'requests.get' and 'BeautifulSoup' parsing are utilized in 'scripts/new_scraper.py', 'examples/after.md', and 'references/api_reference.md'. 2. Boundary markers: The skill does not define specific boundary markers or 'ignore' instructions for the data scraped from external websites. 3. Capability inventory: The skill has capabilities for network requests and file system writes (CSV and generated scripts). 4. Sanitization: While the skill suggests 'Data Cleaning' (Chapter 8), it lacks programmatic sanitization or isolation of scraped content before it is processed by the agent.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 25, 2026, 06:43 AM