configure-ecc

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). The URL points to a GitHub repository hosted on a personal/unknown account (affaan-m) rather than an official vendor or well-known project — while it’s not a direct executable or passworded archive, cloning and installing arbitrary repos from untrusted accounts can include malicious scripts or payloads, so treat it as suspicious absent community trust or verification.

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

• Third-party content exposure detected (high risk: 0.85). Step 0 clones a public GitHub repository at runtime and Step 2/3 copies the included skill/rule directories (including their SKILL.md and other markdown/text files) into the user’s Claude Code context, so the LLM ingests outsider-authored free text from that public repo.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly runs git clone https://github.com/affaan-m/everything-claude-code.git at runtime to fetch the ECC repository, and the fetched files (skills/rules SKILL.md content) are then copied into the agent's skills/rules directories so remote content would directly control agent prompts/behavior and is a required dependency.