insecure-source-code-management

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a methodology guide for authorized security assessments. It provides instructions on how to detect and recover metadata from exposed version control systems and emphasizes that recovered credentials must be treated as sensitive data.
  • [EXTERNAL_DOWNLOADS]: The skill references several open-source security tools hosted on GitHub for data recovery tasks, including arthaud/git-dumper, internetwache/GitTools, WangYihang/GitHacker, anantshri/svn-extractor, sahildhar/mercurial_source_code_dumper, gehaxelt/ds-store, and lijiejie/ds_store_exp. These are standard tools for the described activities.
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection attack surface as it involves the agent ingesting and processing data from external, untrusted web servers.
  • Ingestion points: Data retrieved from probed paths such as /.git/HEAD, /.svn/entries, and .env files (SKILL.md).
  • Boundary markers: The instructions do not define explicit delimiters or 'ignore' instructions for the external content.
  • Capability inventory: The skill allows the agent to perform network requests and use recovery tools to extract data from target servers.
  • Sanitization: There are no mentioned mechanisms for sanitizing or validating the content retrieved from remote endpoints before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 01:23 PM
Security Audit — agent-trust-hub — insecure-source-code-management