hol-guard
Open-source antivirus for AI agents: block risky tools, secret access, prompt injection, malicious packages, MCP servers, plugins, and skills at runtime.
No outside reports on file
Hol-guard reduces risk but is not foolproof—treat it as one layer of defense, not complete protection against determined attacks or zero-day prompt injection techniques.
Based on public adoption and transparency signals as of Oct 26, 2026. Our opinion, not a code audit or security guarantee.
What it can access
- Local filesNo local folder access notedLOW
- Command executionNo shell command runner notedLOW
- Network reachConnects to outside service APINOTED
- CredentialsRequires API key or tokenREVIEW
- Package sourceGitHub (hashgraph-online/hol-guard)PUBLIC
Inferred from this listing’s public package details and description. Check the project’s source before installing.
Install
No npm or PyPI package is listed. Install from source: github.com/hashgraph-online/hol-guard ↗
Everything known
We found nothing about this project in the sources we read (last checked 2026-10-30). That does not mean it has no problems.
0 incidents · 0 advisories · 0 news articles · 0 community threads · checked 2026-10-30
Popularity grade: C (from public signals such as stars and downloads, not a code review). Graded 2026-10-26.
Sources we read: OSV, GitHub Advisory Database, CISA Known Exploited Vulnerabilities, 44 security news feeds and our incident library. Last checked 2026-10-30.
About Hol Guard
Hol-guard is a security filter for AI agents that runs on your machine to block dangerous actions in real time. It catches risky tool calls, attempts to access secrets, prompt injection attacks, malicious code packages, and suspicious AI plugin connections before they can cause harm.
Who it’s for
Developers and organizations running AI agents or Claude Desktop who want to add a safety layer between the AI and sensitive systems. It's for teams concerned about unauthorized tool use, data leaks, or code injection through AI prompts.
Alternatives
- Anthropic's built-in tool restrictions (Claude's native safety guidelines for tool use)
- Sandbox environments (Docker, virtual machines to isolate agent execution)
- API gateways and proxies (control what external services agents can reach)
- Security monitoring tools (CloudTrail, audit logs to detect unauthorized access after the fact)
Setup
Check these steps against the project’s own instructions before running anything.
Follow the project's README on GitHub for installation and configuration instructions specific to your setup.
Why this grade
Signal-based grade (3/7 points): well starred on GitHub, public source code. This reflects adoption and transparency, not a code review.
Popularity and upkeep
- GitHub stars
- 1k
- Listing last refreshed
- 2026-10-30
These are public popularity signals, not a maintenance check. Look at the repository’s commit history and open issues.
Risks to watch for
- Secret exposure: Even with filtering, environment variables or API keys in your system memory could leak if the agent finds unexpected ways to access them
- Prompt injection bypasses: Sophisticated attackers might craft prompts that evade the filter's detection rules
- False sense of security: The tool blocks known risky patterns, but new attack methods may slip through undetected
- Incomplete tool coverage: If your AI uses tools or plugins hol-guard doesn't recognize, they won't be filtered
- Configuration errors: Misconfigured rules could block legitimate work or leave actual threats unblocked
How to lower the risk
Least-privilege permissions
- Run the agent with minimal OS permissions; don't use admin or root accounts
- Restrict which tools and APIs the agent can actually call, separate from hol-guard
Network isolation
- Keep agents on isolated networks if handling very sensitive data
- Block unnecessary outbound connections at the firewall level
Monitoring
- Log and review blocked actions regularly to spot attack patterns
- Alert on repeated block events, which may signal intrusion attempts
Red flags to watch for
- Agent suddenly requesting unusual tool access or secrets it never needed before
- Frequent filter blocks from the same prompt or user
- Configuration changes you didn't make to hol-guard rules
Questions
How do I install Hol Guard?
Hol Guard does not publish a standalone npm or PyPI package. Clone or follow the setup instructions in its source repository at github.com/hashgraph-online/hol-guard.
Are there security reports about Hol Guard?
No reports found for this package in our research (last checked 2026-10-30). Hol Guard currently holds Grade C (Moderate risk signals) based on public adoption and transparency signals. Key risk areas to watch include secret exposure, prompt injection bypasses, false sense of security.
Is Hol Guard actively maintained?
Hol Guard has 1k GitHub stars. Our directory data was last refreshed on 2026-10-30. Check recent commit activity, release tags, and open issues at github.com/hashgraph-online/hol-guard before pinning it in production.
What are alternatives to Hol Guard?
Notable alternatives to Hol Guard include Anthropic's built-in tool restrictions, Sandbox environments, API gateways and proxies, Security monitoring tools. You can also compare Hol Guard side by side with other MCP servers in the Related tools section on this page.
README badge and team notes
Maintain or use Hol Guard? Show its live grade in your GitHub README, docs site, or run a check from the terminal before you install.
GitHub README Markdown:
[](https://agentarmorhq.com/skills/hashgraph-online-hol-guard)HTML embed:
<a href="https://agentarmorhq.com/skills/hashgraph-online-hol-guard"><img src="https://agentarmorhq.com/api/badge/hashgraph-online-hol-guard" alt="AgentArmorHQ Risk Grade for Hol Guard" height="20" /></a>Terminal check:
curl -fsSL https://agentarmorhq.com/api/cli | node --input-type=module - check "hashgraph-online-hol-guard"Share with your security lead
Need sign-off to use hol-guard? Copy this note for Slack, a ticket or a pull request.
### AgentArmorHQ Security Review Memo: hol-guard
- **Listing URL:** https://agentarmorhq.com/skills/hashgraph-online-hol-guard
- **Type:** App connector
- **Risk Grade:** Grade C (Moderate risk signals)
- **Outside Security Advisories on File:** 0
- **Public Adoption:** 1k GitHub stars
#### Permission & Blast-Radius Signals
- **Local files:** No local folder access noted
- **Command execution:** No shell command runner noted
- **Network reach:** Connects to outside service API
- **Credentials:** Requires API key or token
- **Package source:** GitHub (hashgraph-online/hol-guard)
#### Recommended Hardened Setup
- **Source Repo:** https://github.com/hashgraph-online/hol-guard
- **Controls:** Pin an exact version in `.mcp.json`, scope access to a single workspace folder, and use a least-privilege read-only token where supported.
_Based on public adoption and transparency signals as of Oct 26, 2026. Our opinion, not a code audit or security guarantee._Key facts
- Type
- App connector
- Publisher
- hashgraph-online
- Source code
- github.com/hashgraph-online/hol-guard
- Est. context load
- ~900 tokens / turn (~5 tool schemas, 0.45% of 200k context)
- Popularity
- 1k GitHub stars
- Last updated
- 2026-10-30
- Last graded
- 2026-10-26
Community
Discussions for Hol Guard
Ask questions or share configuration tips. Governed by our Community Rules & Terms.
No posts match this filter yet.
Be the first to ask a question or share a working tip.
Check before you install, in Claude
Your assistant answers Block, Review or No known issues for any package before it installs it.
Get this check inside Claude