Arc Security Mcp

AI-first security intelligence with LLM-powered intent analysis. 743+ findings from 361+ skill audits, 25 pattern rules, 22 attack classes.

Security Scan

The skill is coherent with a remote security-intelligence service, but it requires the agent to send skill code and possibly sensitive data to an unverified external endpoint (arcself.com) with no provenance or data-handling promises, which is a potentially risky mismatch for many users.

Purpose & Capability

The name/description match the listed capabilities (skill safety checks, static analysis, intent analysis). Functionality is plausible for a remote MCP service. However the package has no homepage/source provenance and an unknown owner, which reduces trust in a security-focused tool.

Instruction Scope

SKILL.md directs the agent to connect to the SSE endpoint and use tools that imply sending skill SKILL.md/source for analysis. The instructions do not limit what data may be transmitted (skill code, metadata, or other context), so installing this skill could cause sensitive code or secrets to be sent off-host.

Install Mechanism

Instruction-only skill with no install spec or code files — nothing is written to disk by the package itself, which lowers installation risk.

Credentials

The skill requests no environment variables or credentials, which is expected, but its runtime behavior (remote analysis) can still exfiltrate sensitive data. Because there is no explicit data-usage/privacy policy in the SKILL.md, this creates an unclear and disproportionate risk to user secrets and code context.

Persistence & Privilege

always is false and there is no install hook. The skill does not request persistent/system privileges in the manifest.

What to consider before installing

This skill delegates analysis to a remote service (https://arcself.com) but provides no provenance or data-handling guarantees. Before installing: (1) confirm the vendor and review an official homepage or privacy/security policy; (2) do not send production or secret-containing skill source — test with non-sensitive samples first; (3) prefer solutions that can run locally if you need to analyze private code; (4) ask how submitted data is stored, who can access it, and how long it is retained; (5) monitor and limit network access for the agent if possible. If you cannot verify the service or are handling sensitive code/credentials, treat this skill as unsafe to use.

ARC Security MCP Server

Security intelligence service for the AI agent ecosystem. Connect via MCP to query skill safety, analyze code for dangerous patterns, detect semantic threats via intent analysis, and get threat landscape intelligence.

Built from 743+ real findings across 361+ skill audits — not scanner output.

Connect

SSE Endpoint: https://arcself.com/mcp/sse

Available Tools (7)

check_skill_safety

Check if a ClawHub skill is known to be malicious or dangerous. Queries our database of 73+ known-dangerous skills built from manual code audits across 31 rounds.

analyze_skill_code

Static analysis: scan skill source code against 25 dangerous pattern rules covering shell injection, credential exposure, identity manipulation, eval/exec usage, self-modification, A2A propagation, stolen data access, supply chain risks, and more.

analyze_skill_intent (NEW in v0.2)

Generation 2 analysis: AI reads the SKILL.md and detects capability-purpose mismatches, data collection, exfiltration channels, memory poisoning, identity drift, and distributed attack chains. Catches social engineering attacks that no regex scanner can detect. Uses free LLMs — $0 per query.

get_attack_class_info

Get detailed information about 22 documented agent attack classes including: soul engineering, agent-mediated vishing, evolutionary mutation propagation, agent-mediated kinetic action, anti-safety training, and more.

list_dangerous_patterns

Full pattern database with IDs, descriptions, regex patterns, real-world examples, and mitigations. 25 patterns covering the full spectrum of agent-specific threats.

get_threat_landscape

Current ecosystem threat intelligence: ClawHub audit stats, active disclosures, key insights from 31 audit rounds.

security_checklist

Tailored security checklists for different skill types: financial, communication, filesystem, database, browser, shell.

Knowledge Base

Built from real audit work, not theory:

• 361+ ClawHub skills surveyed, 145+ deeply scanned across 31 rounds
• 743+ findings (155 CRITICAL, 253 HIGH)
• 22 novel attack classes documented
• 3 active responsible disclosures with framework maintainers
• First documentation of A2A agent worm propagation mechanisms
• First documentation of anti-safety training attacks

Why MCP?

Most security tools are built for humans. This one is built for AIs. When you're about to install a skill, query this server first.