skills.rest
prompt-security
prompt-securityOFFICIAL
Explore all skills by this author →
GitHub

clawsec-nanoclaw

Checks NanoClaw skills against a signed security advisory feed and verifies package signatures.

1.1k|113|Updated Feb 5, 2026
Search Tags:#Security#vulnerability-scanning#MCP-tools#signature-verification#nanoClaw#advisory-feed#integrity-monitoring
One-click install
npx skills add https://github.com/prompt-security/clawsec --skill clawsec-nanoclaw
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: clawsec-nanoclaw
Source: https://github.com/prompt-security/clawsec/tree/main/skills/clawsec-nanoclaw
Command: npx skills add https://github.com/prompt-security/clawsec --skill clawsec-nanoclaw

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

NanoClaw agents can install skills with known vulnerabilities or tampered packages without any warning. This Skill scans installed skills against a curated, Ed25519-signed advisory feed, blocks unsafe installations, verifies package signatures, and monitors critical configuration files for unauthorized changes.

Core Features & Use Cases

  • Advisory Monitoring: Fetches and caches the signed ClawSec advisory feed (NVD CVEs, community advisories, GHSA records) every 6 hours with fail-closed Ed25519 signature verification.
  • Pre-Installation Safety Checks: The clawsec_check_skill_safety MCP tool reports known vulnerabilities and exploitability scores before a skill is installed.
  • Package Signature Verification: clawsec_verify_skill_package validates Ed25519 signatures on skill packages using a pinned public key and bounded path policy.
  • File Integrity Monitoring: Detects drift in protected files like registered_groups.json and CLAUDE.md, auto-restores critical files, and maintains a hash-chained audit log.
  • Use Case: Before installing a new community skill, the agent runs a safety check, finds a critical advisory with high exploitability, and warns the user instead of installing it.

Quick Start

Ask your NanoClaw agent to check whether any installed skills have known security advisories before installing a new skill.

Frequently Asked Questions about clawsec-nanoclaw

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I check NanoClaw skills for security vulnerabilities?▼

Use the clawsec_check_advisories MCP tool to scan all installed skills against the ClawSec advisory feed. It returns matching advisories with severity and exploitability scores so you can prioritize critical or highly exploitable issues first.

How do I verify a skill package signature before installation?▼

Call clawsec_verify_skill_package with the absolute path to the package file. It verifies the Ed25519 signature against the pinned ClawSec public key and returns an install, block, or review recommendation. Package paths must be under approved directories like /tmp or /workspace/project/data.

Does ClawSec work with NanoClaw container deployments?▼

Yes, it is designed for NanoClaw's container architecture. MCP tools run inside the container while the advisory cache manager and signature verification run on the host, communicating through IPC handlers registered in src/ipc.ts.

How often does the advisory feed update?▼

The host-side AdvisoryCacheManager fetches the signed feed from clawsec.prompt.security every 6 hours by default. You can adjust the refresh interval in the host startup code or trigger an immediate refresh with the clawsec_refresh_cache tool.

Why does signature verification fail on a downloaded skill package?▼

Verification fails if the package was tampered with, the signature file is missing, or the file path is outside the allowed directories. Re-download both the package and its .sig file, use absolute paths, and confirm the pinned public key file exists in the skill's advisories directory.

What files does NanoClaw integrity monitoring protect?▼

It protects registered_groups.json and CLAUDE.md files as critical priority with automatic restore, plus container and host TypeScript code as alert-only targets. Baselines use SHA-256 hashes and all operations are recorded in a hash-chained audit log.

Related Skills

Explore More Software Engineering Skills→
chiptoe-svgchiptoe-svg

x-integration

Automate X interactions via WhatsApp using Playwright browser automation.

Community
Intermediate
noname300989noname300989

cron

Manage gateway job scheduling with cron syntax for recurring and one-shot tasks.

Community
Intermediate
SystemTceSystemTce
33

dify-security-deployment

Configure SSRF protection, sandbox isolation, and observability for Dify deployments.

Community
Advanced
MolcajeteAIMolcajeteAI
2

dependency-security

Scan npm dependencies for vulnerabilities using npm audit, Snyk, and Socket.dev.

Official
Intermediate