1. X
  2. HOL
Log inSign up
HOL
5,849 posts
Image
user avatar
HOL
@HashgraphOnline
A neutral home for open AI-agent infrastructure. Standards, registries, discovery, trust, and coordination across Web2 + Web3.
hol.org
Joined March 2024
23
Following
16.2K
Followers
RepliesRepliesMediaMedia

New to X?

Sign up now to get your own personalized timeline!

Create account

By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.

Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
Don't miss what's happening
People on X are the first to know.
Log inSign up
  • Pinned
    user avatar
    HOL
    @HashgraphOnline
    May 4
    1/ Today we’re launching the HOL Partner Program. Cohort One brings together 30+ signed partners, including XMTP, GoDaddy, and DSR, to help shape open infrastructure for AI agents. Registries. Payments. Privacy. Security. Communication. Standards. The agent stack is forming
    Image
    00:00
    3.1M03.1M
  • user avatar
    HOL
    @HashgraphOnline
    5h
    axios GHSA-xj6q-8x83-jv6g: prototype pollution can inject Basic auth. Call axios with an auth object missing username/password when Object.prototype is polluted, and axios builds the Authorization header from attacker-controlled values.
    Image
    AI Supply Chain Security — Package, Skill & Tool Risks | HOL Guard
    From hol.org
    5450545
    user avatar
    HOL
    @HashgraphOnline
    5h
    axios ships as a transitive dep in most MCP servers and HTTP tools. HOL Guard keeps an AIBOM of every component your agent loads and scans npm installs locally, so you see instantly whether axios is present and at what version. curl -fsSL hol.org/guard/install.… | bash
    56056
  • user avatar
    HOL
    @HashgraphOnline
    7h
    Your AI coding agent has shell access. It can read your .env, install packages, and push to prod — same permissions as you, none of the judgment. There's no safety layer between "vibes-based approval" and your filesystem. We built one. HOL Guard: open-source antivirus for AI
    Image
    HOL Guard — Stop AI agents from running risky tools
    From hol.org
    6620662
    user avatar
    HOL
    @HashgraphOnline
    7h
    Replying to @HashgraphOnline
    The market context: • Uber: 5K engineers on Claude Code. Blew their 2026 AI budget in 4 months. 11% of deploys are now agent-only, no human in the loop. • NVIDIA: 30K devs on Cursor. 3x more committed code. • Stripe: 4K+ machines running AI agents. Only 43% of orgs have
    1000100
    user avatar
    HOL
    @HashgraphOnline
    7h
    Install in 60 seconds: hol.org/guard Open source (Apache 2.0): github.com/hashgraph-onli… Free tier: full local protection forever. Pro: $15/mo for cloud sync. Team: $29/seat/mo for fleet dashboard & policies.
    Image
    HOL Guard — Stop AI agents from running risky tools
    From hol.org
    65065
  • user avatar
    HOL
    @HashgraphOnline
    10h
    The upload_ad_image tool in meta-ads-mcp (PyPI, < 1.0.115) takes an attacker-controlled image_url and passes it straight into httpx with follow_redirects=True and no scheme, host, or IP validation. In streamable-http transport that means an unauthenticated remote attacker gets
    Image
    Data overexposure | HOL Guard Security
    From hol.org
    5800580
    user avatar
    HOL
    @HashgraphOnline
    10h
    HOL Guard intercepts MCP tool calls before they run, so when an agent tries to invoke upload_ad_image with an image_url pointing at 169.254.169.254 or an internal host, it gets flagged and held for approval instead of silently fetched. It also scans the MCP server itself at load
    Image
    HOL Guard — Stop AI agents from running risky tools
    From hol.org
    79079
  • user avatar
    HOL
    @HashgraphOnline
    13h
    if you run AI agents that call tools, HOL Guard sits in the loop and enforces what each call is allowed to do: policy bundles, tool-call approval, MCP servers scanned at load time against an AIBOM of installed components.
    8350835
    user avatar
    HOL
    @HashgraphOnline
    13h
    Replying to @HashgraphOnline
    uv tool install hol-guard==2.0.1127
    56056
    user avatar
    HOL
    @HashgraphOnline
    13h
    Changelog: github.com/hashgraph-onli… Docs: hol.org/guard
    57057
  • HOL reposted
    user avatar
    Aaliyah
    @Allys_web3
    13h
    Your code is your competitive advantage. Why trust it to a security tool that demands mandatory cloud syncing? HOL Guard is built from the ground up to respect developer privacy: *Open-Source Runtime: Fully transparent codebase available right on GitHub. *Local-First
    Image
    2170217
  • HOL reposted
    user avatar
    kimcĦi.ℏ/acc
    @HederaKimchi
    Jul 17
    🤖 Hashgraph Online Builds Open Standards for AI Agents @HashgraphOnline (HOL) is developing open-source standards and tools that enable AI agents to identify, verify, discover, and communicate across Web2 and Web3 using the @hedera Consensus Service (HCS). 📊 Current
    Image
    user avatar
    kimcĦi.ℏ/acc
    @HederaKimchi
    Jul 17
    Article cover image
    Article
    🤖 Hashgraph Online Builds Open Standards for the AI Agent Era
    @HashgraphOnline (HOL) is building open-source standards and developer tools that enable AI agents to identify one another, discover services, and communicate securely. HOL is an open-source...
    1.5K01.5K
  • user avatar
    HOL
    @HashgraphOnline
    Jul 17
    If you run AI agents with tool access, HOL Guard sits between the agent and your system and audits every tool call against policy, so a misbehaving prompt can't touch what it shouldn't. v2.0.1113 fixes the runtime incorrectly blocking routine GitHub mutations. Agents doing
    2.8K02.8K
    user avatar
    HOL
    @HashgraphOnline
    Jul 17
    Full release notes: github.com/hashgraph-onli… Docs and quickstart: hol.org/guard
    Image
    Release v2.0.1113 · hashgraph-online/hol-guard
    From github.com
    2500250
  • HOL reposted
    user avatar
    FatBoy
    @Shootitfatboy
    Jul 17
    If your vibe coding or actually coding and using Ai you need this. Fullstop
    user avatar
    HOL
    @HashgraphOnline
    Jul 17
    If you give AI agents tool access, HOL Guard sits between the agent and your system, auditing every tool call and enforcing approval rules before anything runs. Across v2.0.1097 to v2.0.1112, command coverage grew substantially and several bypass paths got closed. New coverage:
    7030703
  • user avatar
    HOL
    @HashgraphOnline
    Jul 17
    If you give AI agents tool access, HOL Guard sits between the agent and your system, auditing every tool call and enforcing approval rules before anything runs. Across v2.0.1097 to v2.0.1112, command coverage grew substantially and several bypass paths got closed. New coverage:
    2.1K02.1K
    user avatar
    HOL
    @HashgraphOnline
    Jul 17
    Changelog: github.com/hashgraph-onli… Docs: hol.org/guard
    2170217
Advertisement
Advertisement