neuronwriter-api

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the agent to ask the user for an API key, use the X-API-KEY header for requests, and log "Commands Used" (and run/emit commands), which forces the LLM to receive and/or include the secret verbatim in generated requests/commands — creating a direct exfiltration risk.

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

• Third-party content exposure detected (high risk: 1.00). The skill repeatedly calls the NeuronWriter API (e.g., new-query / get-query via scripts/neuronwriter.sh) and explicitly ingests fields like serp_summary, competitors, ideas.*, and terms_txt.content_basic_w_ranges (derived from public competitor pages and SERP data) and then requires the agent to parse and enforce those recommendations to generate, import, and iterate content — so untrusted/public third‑party content is both read and used to drive decisions and actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill makes runtime API calls to https://app.neuronwriter.com/neuron-api/0.5/writer to fetch term ranges, SERP intent, and other recommendations that are directly injected into prompts and used to control content-generation and import/score loops, making this external endpoint a required runtime dependency that controls agent instructions.