skills/hk-hub/agentskills/paper-fetch/Gen Agent Trust Hub

paper-fetch

Pass

Audited by Gen Agent Trust Hub on Aug 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill is designed to download PDF files from various academic repositories and search APIs, including Unpaywall, Semantic Scholar, arXiv, PubMed Central, and bioRxiv.
  • [COMMAND_EXECUTION]: The script uses subprocess.run to invoke a companion script (cloak_pdf.py) for handling browser-based downloads and to check for the presence of the cloakbrowser Python dependency.
  • [SAFE]: Implements extensive Server-Side Request Forgery (SSRF) protection. It validates all URLs and resolved IP addresses against a blocklist that includes loopback aliases, private IP ranges (RFC 1918), and cloud metadata endpoints (Google, AWS) in both the main script and its companion.
  • [SAFE]: Enforces security checks on all downloaded content, verifying the %PDF magic bytes at the start of the file and imposing a 50 MB size limit to prevent resource exhaustion attacks.
  • [SAFE]: Uses shlex.quote when generating suggested follow-up commands in the output, preventing potential command injection if a user-provided DOI contains malicious shell characters.
  • [SAFE]: Browser automation for bypassing Cloudflare (via CloakBrowser) and institutional access modes are strictly opt-in via environment variables set by the operator, maintaining a clear trust boundary.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 19, 2026, 05:30 PM
Security Audit — agent-trust-hub — paper-fetch