secrets-in-git-history

Pass

Audited by Gen Agent Trust Hub on Aug 3, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill is designed to process untrusted data from external Git repositories (history, logs, and files), which constitutes an attack surface for indirect prompt injection.
  • Ingestion points: Data is ingested from repository clones and GitHub/GitLab API responses as described in SKILL.md and reference/git-archaeology.md.
  • Capability inventory: The skill utilizes shell-based tools including git, curl, jq, trufflehog, and gitleaks.
  • Sanitization: No explicit content sanitization or validation of the processed Git data is documented.
  • Boundary markers: The skill does not employ specific delimiters to isolate external data from the agent's instructions.
  • [DATA_EXFILTRATION]: Performs network requests to GitHub's official API (api.github.com) and commit patch endpoints to resolve user identities and extract metadata.
  • [COMMAND_EXECUTION]: Provides a comprehensive set of shell commands for local Git repository analysis, including the use of the Git pickaxe (-S), tree searching (git grep), and execution of specialized security scanners like TruffleHog and Gitleaks.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 3, 2026, 08:31 PM
Security Audit — agent-trust-hub — secrets-in-git-history