Automation & Ops security-auditsecurity-advisoriesnanoclawmcpskill-integritysupply-chain-security

NanoClaw Security Sentinel

Check NanoClaw skills for known risks before installation and during operation.

FollowSkills review · FSRS-2.0
Not recommended
44/ 100 5-point scale 2.2 / 5
Trust14 / 25 · 2.8/5

The evidence shows host-side IPC, Ed25519 verification, a pinned public key, bounded paths, symlink rejection, atomic cache writes, container isolation for baselines, and quarantine/restore handling; remote-feed data flow is disclosed. Points are deducted because automatic restoration is enabled by default and changes critical files, user confirmation is not enforced, stale cache use is possible, and publisher identity plus the key-establishment path remain incomplete.

Reliability7 / 20 · 1.8/5

The documentation covers installation, integration, troubleshooting, and failure logging; the code includes timeouts, fail-closed signature checks, atomic writes, and refresh de-duplication. Points are deducted because this is a static review capped at 10, version-range matching is still marked TODO despite the changelog claiming support, and several import, path, and runtime-environment assumptions are not reproducible from the selected evidence.

Adaptability5 / 15 · 1.7/5

The audience, NanoClaw use cases, triggers, and non-fit cases are mostly clear, with several MCP tools. Points are deducted for no Chinese-language support or localized output and reliance on a single overseas HTTPS advisory service; mainland-China reachability and offline behavior are not sufficiently bounded despite cache fallback documentation.

Convention9 / 15 · 3.0/5

The directory provides layered installation, configuration, examples, FAQs, troubleshooting, limitations, uninstall instructions, AGPL-3.0-or-later licensing, versioning, a changelog, and support/issue channels. Points are deducted because maintenance ownership is broad rather than operationally explicit, updates depend on external services, and some documented versions, paths, and implementation details conflict.

Effectiveness5 / 15 · 1.7/5

The claimed functions can plausibly perform advisory checks, package verification, and file-integrity monitoring, with directly usable MCP examples. Points are deducted because static evidence cannot establish end-to-end operation; version matching remains unfinished, deployment requires host integration and an external feed, and outputs plus automatic restoration still require human review.

Verifiability4 / 10 · 2.0/5

Committed implementation code, type definitions, a pinned key, policy, changelog, and concrete interface examples provide limited auditability. Points are deducted because no execution occurred, the selected evidence lacks verifiable CI and comprehensive key-path tests, and claims such as “zero false positives” and “production-ready” are unsupported.

Evidence confidence:Low Reviewed Jul 20, 2026 Reviewed revision eac8392f8ac9
Before you use it
  • Before installation, verify the NanoClaw and Node.js versions, host paths, and every MCP/IPC integration point; the documented .js imports may require a build step for the provided .ts sources.
  • Do not treat a valid signature or absence of advisories as proof of complete safety; compromised signing keys, zero-days, malicious signed packages, and post-check modifications remain possible.
  • Use the default automatic restoration cautiously in production, preserving quarantine, patch, and audit artifacts; ensure stale cache data is not presented as a current security conclusion.
  • Version-range matching is marked TODO and should be checked before relying on version-specific advisory results.
  • The core feed depends on an overseas domain; verify mainland-China connectivity and prepare a trusted mirror or explicit offline policy.
See the full review method →

What does this skill do, and when should you use it?

ClawSec for NanoClaw monitors security advisories for NanoClaw skills and dependencies used by WhatsApp bots. Its MCP tools check prospective and installed skills, expose exploitability context, and support package signature, file-integrity, baseline, and audit-log verification. It is a fit for teams that need supply-chain and advisory monitoring around NanoClaw. It is explicitly not intended for code review, performance troubleshooting, or general debugging.

It calls clawsec_check_skill_safety for a prospective skill, clawsec_check_advisories for an installed-skill audit, and clawsec_list_advisories to filter the advisory feed by severity, type, or exploitability. It also exposes clawsec_verify_skill_package, clawsec_refresh_cache, clawsec_check_integrity, clawsec_approve_change, clawsec_integrity_status, and clawsec_verify_audit. The skill consumes the signed feed at https://clawsec.prompt.security/advisories/feed.json and uses /workspace/project/data/clawsec-advisory-cache.json as its stated cache location.

  1. A NanoClaw administrator checks a requested skill before installing it.
  2. A bot maintainer answers whether the installed skills have known vulnerabilities.
  3. A security operator investigates suspicious skill behavior using advisory and exploitability context.
  4. An operations team runs recurring audits and alerts on critical or highly exploitable matches.
  5. A standalone installer verifies the release manifest, signature, public-key fingerprint, archive, and package files before extraction.

What are this skill's strengths and limitations?

Pros
  • Covers pre-install checks, installed-skill audits, and advisory browsing.
  • Combines severity with exploitability context for triage.
  • Includes package verification and file-integrity-related operations.
  • Targets NanoClaw directly and provides a daily scheduled-check example.
Limitations
  • Requires network access and an MCP-capable runtime.
  • The supplied source omits the detailed INSTALL.md and docs content, so deployment configuration requires consulting the repository.
  • It does not cover code review, performance issues, or general debugging.
  • The stated default audit and cache paths may need adjustment for a specific environment.

How do you install this skill?

Install it with the Vercel Skills CLI: npx skills add prompt-security/clawsec --skill clawsec-nanoclaw -a openclaw -y. The supplied material does not include the full manual NanoClaw deployment procedure; for standalone releases, follow the provided manifest-verification procedure before extracting or installing the archive.

How do you use this skill?

Before installing a skill, call clawsec_check_skill_safety({skillName: "new-skill", skillVersion: "1.0.0"}). To audit installed skills, call clawsec_check_advisories({installRoot: "/home/node/.claude/skills"}) and prioritize matches with critical severity or high exploitability. Example user triggers include “Are my skills secure?” and “Check my NanoClaw skills for security advisories.”

FAQ

Is this skill free?
The supplied material does not state a service pricing model. The repository identifies its license as AGPL-3.0.
What data does it access?
It reads installed skills and their installation directory and accesses the signed advisory feed; the exact permissions depend on the MCP tools and host runtime.
Does it automatically block risky installations?
The example flow displays risks and requests confirmation. The supplied material does not establish that every installation path is forcibly blocked.
How often is the advisory feed updated?
The SKILL.md states that it updates automatically every six hours.

More skills from this repository

All from prompt-security/clawsec

Automation & Ops

NanoClaw Traffic Guardian

A host-managed baseline for optional NanoClaw traffic inspection and runtime security findings.

Automation & Ops

Picoclaw Security Guardian

A read-first security layer for Picoclaw advisory checks, configuration drift detection, and signed release verification.

Automation & Ops

ClawSec Suite Manager

Protect OpenClaw skill installs with signed advisories, risk checks, and approval-gated responses.

Automation & Ops

Picoclaw Security Posture Review

Review Picoclaw security posture in read-only mode and give operators safe remediation guidance.

Automation & Ops

ClawSec Release Manager

A maintainer workflow for versioning, releasing, and verifying Claw skills through GitHub.

Automation & Ops

ClawSec Advisory Feed

Keep OpenClaw agents informed about emerging security threats.

Automation & Ops

OpenClaw Audit Watchdog

Automates daily OpenClaw security audits and delivers actionable reports.

Automation & Ops

ClawHub Reputation Gate

Checks ClawHub reputation before OpenClaw skill installation and requires explicit approval for low-scoring skills.

Automation & Ops

Soul Guardian File Integrity Guard

Detect and handle unauthorized changes to an agent’s core workspace files.

Automation & Ops

Hermes Attestation Guardian

Generate verifiable Hermes security attestations and detect authenticated baseline drift.

Automation & Ops

Clawtributor Security Reporter

Prepare sanitized AI-agent incident reports for explicit human approval before submission.

Automation & Ops

ClawSec Scanner

Unify dependency, CVE, source-code, and OpenClaw hook risk scanning in one agent-security workflow.

Automation & Ops

Picoclaw Traffic Guardian

An opt-in, detect-and-log baseline for monitoring Picoclaw gateway traffic and exporting security posture data.

Automation & Ops

OpenClaw Traffic Guardian

A scoped, opt-in runtime traffic monitoring baseline for OpenClaw.

Automation & Ops

Hermes Traffic Guardian

A consent-based, auditable baseline for monitoring Hermes runtime traffic.

Related skills