bug-bounty

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (critical risk: 1.00). This repository contains explicit, operational instructions and tooling to perform real-world offensive actions (credential-spraying orchestration, SSRF to cloud metadata for key exfiltration, exploit chaining, and autonomous attack loops) which clearly enables harmful activity beyond benign explanation.

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

• Third-party content exposure detected (medium risk: 0.30). Validator Agent at agents/validator.md ingests the externally-authored “finding” description you provide (and optionally Burp proxy history responses), then uses it to decide PASS/KILL/DOWNGRADE, so outsider free text can reach the LLM without selecting a specific pre-defined item.

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

• Potentially malicious external URL detected (high risk: 0.90). The repo and installer URLs are invoked in the project install instructions and will fetch and (in the case of the Ollama installer via curl | sh) execute remote code — e.g., git clone https://github.com/shuvonsec/claude-bug-bounty.git and curl -fsSL https://ollama.ai/install.sh | sh are present in the docs/installation steps.