Back to skill
Skillv1.0.3
ClawScan security
Reward Sciences · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignApr 22, 2026, 11:49 AM
- Verdict
- benign
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- The skill's instructions, required actions, and external dependencies are consistent with its stated purpose: it uses the Membrane CLI to connect to Reward Sciences and does not ask for unrelated credentials or system access — but it does require trusting the Membrane service and installing a third-party CLI tool.
- Guidance
- This skill appears coherent and does what it says: it relies on the Membrane CLI to broker access to Reward Sciences. Before installing/using it, consider: 1) Vet the @membranehq/cli package and its publisher (npm page, repo, recent activity, reviews) before installing globally; use the provided npx examples if you prefer not to install globally. 2) Understand that you are delegating credential management to Membrane — review their privacy/security practices and terms. 3) In headless environments you'll need to perform the OAuth browser step (obtain and paste the code) — avoid pasting codes from untrusted sources. 4) Because the skill requires network access and a Membrane account, ensure your environment/network policies permit that outbound access. If you need more assurance, provide the package name and repository URL and verify its source code before installing.
Review Dimensions
- Purpose & Capability
- okName/description match the runtime instructions: the SKILL.md describes interacting with Reward Sciences via the Membrane CLI, creating connections, listing/creating actions, and running them. Nothing requested (no env vars, no config paths) is unrelated to this purpose.
- Instruction Scope
- okInstructions are scoped to using the Membrane CLI (login, connect, action list/create/run). They do not instruct reading arbitrary local files, exporting unexpected data, or accessing unrelated credentials. Headless login flow requires the user to paste a code, which is normal for CLI OAuth flows.
- Install Mechanism
- noteThe SKILL.md recommends installing @membranehq/cli globally via npm (or using npx in one example). There is no install spec in the manifest (instruction-only skill). Installing a global npm package writes to the system and requires trusting that package/publisher; this is expected for a CLI-based integration but users should vet the package and prefer npx when possible.
- Credentials
- okThe skill declares no required environment variables or credentials and explicitly advises not to ask users for API keys (Membrane manages auth server-side). The only external requirement is a Membrane account and network access, which is proportionate to the described functionality.
- Persistence & Privilege
- okalways is false and there are no instructions to modify other skills or global agent configs. The skill does not request persistent system privileges beyond using the Membrane CLI when installed by the user.