cf-crawl

Warn

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses eval to export environment variables sourced from local configuration files. This pattern is vulnerable to command injection if an attacker can control the contents of the .env or ~/.env files.
  • [DATA_EXFILTRATION]: The workflow attempts to access sensitive files including .env and ~/.env to retrieve API credentials. Accessing configuration files in the home directory is a high-risk behavior as these files often contain secrets for multiple services.
  • [EXTERNAL_DOWNLOADS]: Fetches content from Cloudflare's official API (api.cloudflare.com). While Cloudflare is a well-known service, this involves the transmission of sensitive API tokens.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8). Ingestion points: External web content fetched via urllib.request.urlopen. Boundary markers: None; content is saved directly to markdown files. Capability inventory: File system write access (open(filepath, 'w')). Sanitization: Absent; raw markdown content from the crawl is written to disk without filtering.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 13, 2026, 04:45 AM