secrets-in-git-history

Pass

Audited by Gen Agent Trust Hub on Aug 24, 2026

Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill documents and targets numerous sensitive file paths and credential formats for discovery and potential extraction.
  • Evidence: high-signal-indicators.md lists files such as .env, terraform.tfstate, id_rsa, and various private key formats.
  • [COMMAND_EXECUTION]: Orchestrates the use of shell commands and third-party security tools to process repository data.
  • Evidence: Uses git, curl, trufflehog, and gitleaks to clone, fetch, and scan repositories.
  • [PROMPT_INJECTION]: The skill ingests and processes untrusted external content from git repositories and APIs, creating a surface for indirect prompt injection.
  • Ingestion points: Remote git repositories (git clone) and GitHub/GitLab API responses (curl) in SKILL.md and reference/git-archaeology.md.
  • Boundary markers: Absent.
  • Capability inventory: Shell command execution (git, curl), security tool execution (trufflehog, gitleaks), and arbitrary file reading (git cat-file, git show).
  • Sanitization: Absent; the skill relies on the agent to interpret raw repository content.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 24, 2026, 11:50 AM
Security Audit — agent-trust-hub — secrets-in-git-history