Skillv1.1.0

ClawScan security

Node.js Project Architecture · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignMar 6, 2026, 6:02 PM

Verdict: benign
Confidence: medium
Model: gpt-5-mini
Summary: An instruction-only architecture guide whose requirements and actions match its stated purpose, but it prescribes patterns (hot-reload admin endpoints and storing admin credentials in config.json) that can be insecure if implemented or deployed without care.
Guidance: This is a coherent, instruction-only architecture guide — it won't itself install code or ask for credentials. However, follow-up considerations before adopting its patterns: - Review any produced admin endpoint code carefully. Ensure requireAdmin is implemented with robust authentication (not a plain header in production), use TLS, enforce rate limits, CSRF protection, and logging. - Avoid storing sensitive secrets (admin passwords, API secrets) in a web-served config.json; prefer environment variables or a secrets manager for production. If you must use config files, ensure /api/config never returns secret fields and that backup files are protected. - Validate and sanitize incoming admin POST data before writing config.json, and consider access controls (IP allowlist, OAuth, or token-based auth) rather than a simple header password. - Treat hot-reload in production cautiously: it simplifies operations but can enable misconfiguration or privilege escalation if exposed publicly. - Because the skill is instruction-only, the security risk comes from code you or an agent generate from these instructions — audit generated server/admin code before deploying publicly.

Review Dimensions

Purpose & Capability: okName/description (AI-friendly Node.js project architecture, file-splitting, config externalization, admin dashboard) match the SKILL.md and reference files. The skill requests no unrelated binaries, env vars, or installs — everything in the docs is coherent with structuring projects and adding an admin UI.
Instruction Scope: noteRuntime instructions focus on splitting files, externalizing config, and adding admin endpoints (/api/config, /admin/config). These are within scope, but the docs explicitly instruct reading and overwriting ./config.json, backing it up, and exposing a config API. That pattern can accidentally expose secrets (if strip logic is incomplete) or enable unauthenticated config changes if 'requireAdmin' is not implemented securely. The guidance gives the agent broad discretion to create network-accessible admin endpoints and perform file writes — expected for this purpose but security-sensitive.
Install Mechanism: okNo install spec and no code files executed by the platform — instruction-only. This minimizes immediate filesystem or network risk from the skill package itself.
Credentials: noteThe skill declares no required environment variables or credentials (consistent). However, it recommends storing admin credentials and third-party API keys in config.json (examples show admin.password and thirdParty keys). Storing secrets in a writable JSON served by the app increases the risk of accidental exposure; using environment variables or dedicated secret storage is safer for production.
Persistence & Privilege: okSkill metadata does not request always:true or elevated privileges and is user-invocable only. The README claim that the skill "auto-activates" on certain prompts is a behavioral description, not a metadata privilege — metadata indicates normal, non-forced inclusion.