git-hooks

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.80). The prompt includes explicit "SELF-EVOLUTION TELEMETRY" instructions to look for a project memory directory in ~/.claude/projects/ and append execution metadata, which is outside the stated purpose of setting up git hooks and thus constitutes a hidden/deceptive side-effect (telemetry/exfiltration) beyond the advertised functionality.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). This skill creates and uses public third‑party hook references (e.g., the .pre-commit-config.yaml pointing to https://github.com/astral-sh/ruff-pre-commit and similar repos) and explicitly runs/install commands like pre-commit install / pre-commit run and package installs that fetch and execute untrusted GitHub/npm content whose outputs the agent reads to diagnose and re-generate artifacts, allowing indirect prompt injection.