researchclaw

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The config wizard explicitly asks for "your API key" (or an env var) and then generates and displays config.yaml and runs connectivity checks, which allows the agent to embed and output API keys verbatim and use them in test requests.

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

• Third-party content exposure detected (high risk: 0.90). The skill's pipeline explicitly performs literature search and ingestion from public third‑party sources (arXiv and Semantic Scholar) in Stage 3/4 (references/pipeline-stages.md and SKILL.md) and uses those summaries for downstream decisions (hypothesis generation, experiment design, paper writing and citation verification), exposing the agent to untrusted external content that could carry indirect prompt injections.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's setup instructions explicitly fetch and install remote code from https://github.com/aiming-lab/AutoResearchClaw (via git clone and pip install), which would pull and execute third‑party code at runtime and is a required dependency for the skill.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (medium risk: 0.60). The skill includes explicit privileged operations (e.g., "sudo apt-get install texlive-full" and "sudo usermod -aG docker $USER") and offers to perform automatable fixes with user approval, which would modify system-level state and require elevated privileges.