js-analyzer

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs hunting for hardcoded secrets (with concrete secret-pattern regexes and tooling) and tells the agent to "document each finding" and include evidence/code snippets, which would require the LLM to handle and potentially output secret values verbatim.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.90). Multiple URLs point to attacker-controlled or typosquatted domains, direct malicious script endpoints, and open-redirect/JSONP payloads that can be used to host or deliver malware or enable client-side attacks, so these are high-risk indicators.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The content contains explicit, actionable instructions and code samples that enable data exfiltration, remote code execution (eval/Function), credential/secret harvesting, and supply-chain/remote-injection techniques — high abuse potential despite pentest framing.

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

• Third-party content exposure detected (medium risk: 0.30). In SKILL.md Phase 1.3/2.2/3.1/4.x the required runtime workflow downloads and parses JavaScript and source maps from a user-supplied target URL (e.g., via requests.get/curl to extract script srcs and then fetch .js/.map), so outsider-authored page/JS content can be ingested for analysis.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill contains explicit runtime instructions to clone and execute remote GitHub repositories (e.g., "https://github.com/m4ll0k/SecretFinder", "https://github.com/GerbenJavado/LinkFinder", "https://github.com/kleiton0x00/ppmap"), which would fetch and run third‑party code during the analysis workflow.