blog-post-counter

Fail

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill instructions in SKILL.md (Step 5) pipe data from a remote URL directly into a python3 process using the -c flag. While the script logic is defined within the skill itself for the purpose of parsing lastmod dates, this pattern allows untrusted remote content to be processed by a high-privilege interpreter.
  • [EXTERNAL_DOWNLOADS]: The skill dynamically fetches robots.txt and sitemap.xml files from arbitrary user-provided domains and integrates with third-party APIs including Ahrefs and DataForSEO.
  • [COMMAND_EXECUTION]: The skill relies on various shell utilities, such as curl, grep, and wc, combined with bash piping to automate web data extraction and post counting.
  • [PROMPT_INJECTION]: The skill ingests untrusted metadata and sitemap content from external websites without boundary markers or sanitization, creating an indirect prompt injection surface where a malicious site could attempt to influence the agent's logic or output during an audit.
Recommendations
  • HIGH: Downloads and executes remote code from: https://example.com/sitemap.xml - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Jun 13, 2026, 08:44 AM
Security Audit — agent-trust-hub — blog-post-counter