Skillv1.9.3

ClawScan security

Argus Intelligence · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignMar 30, 2026, 3:48 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's declared purpose (on‑chain intelligence and risk scanning) aligns with its runtime instructions and requirements: it only needs curl plus an ARGUS_ENDPOINT and uses that endpoint to call remote APIs — nothing requested appears disproportionate or hidden.
Guidance: This skill appears internally consistent, but before installing: 1) Verify ARGUS_ENDPOINT is the official https://argus.getfailsafe.com (check TLS cert and homepage) to avoid pointing your agent to an attacker-controlled host. 2) Do not embed or expose secret payment tokens to endpoints you haven't verified; prefer short‑lived tokens and treat X-Stripe-Token/X-Payment-Proof as sensitive. 3) Only register webhooks to endpoints you control; store the returned webhook secret securely and rotate it if leaked. 4) Understand that agentId and webhook URLs you provide will be sent to the remote service. 5) Optionally review the linked GitHub repo and privacy/terms on getfailsafe.com for additional trust signals before use.
Findings: [no_scan_findings] expected: The regex-based scanner produced no findings. This is expected because the skill is instruction-only (SKILL.md) and has no executable code files for static analysis.

Review Dimensions

Purpose & Capability: okName/description describe onchain intel, and the skill only requires curl and ARGUS_ENDPOINT to call remote ARGUS API endpoints (token analysis, address risk, prompt-check, webhooks, etc.). These requirements are appropriate for the stated purpose.
Instruction Scope: noteSKILL.md instructs the agent to send requests (including agentId) to the configured ARGUS_ENDPOINT, to register webhooks (and store the returned webhook secret), and to optionally supply payment headers (X-Stripe-Token / X-Payment-Proof). These actions are consistent with a remote SaaS integration, but they do involve sharing agentId and receiving/storing a webhook secret — the user should ensure webhook endpoints and any tokens are handled securely and that the ARGUS_ENDPOINT is legitimate.
Install Mechanism: okInstruction-only skill with no install spec or extracted downloads. No code is written to disk by an installer. This is low risk and appropriate for a wrapper that documents curl usage.
Credentials: okOnly ARGUS_ENDPOINT (and the curl binary) are required. ARGUS_ENDPOINT is the service endpoint and not a secret by itself; payment tokens are optional and supplied per-request. The requested env var is proportional to the skill's purpose. Note: because ARGUS_ENDPOINT controls where requests go, ensure it points to the official service before installation.
Persistence & Privilege: okSkill is not forced-always, does not request any special agent-wide privileges, and contains no instructions to modify other skills or system-wide agent settings. Autonomous invocation is allowed (platform default) and is expected for this kind of integration.