recon-bakeries

Pass

Audited by Gen Agent Trust Hub on Jul 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses standard command-line utilities including curl, jq, grep, and sort to perform its intended reconnaissance tasks against target domains.
  • [EXTERNAL_DOWNLOADS]: Employs curl to communicate with target web servers to perform platform fingerprinting and test the accessibility of specific e-commerce API endpoints.
  • [PROMPT_INJECTION]: The skill ingests untrusted data from target web responses via curl (SKILL.md) and processes it with local tools. While it lacks explicit boundary markers or sanitization, this vulnerability surface is characteristic of web reconnaissance tools and represents a low risk in this context.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 9, 2026, 05:19 PM
Security Audit — agent-trust-hub — recon-bakeries