Skip to content
View glatinone's full-sized avatar
🏠
Working from home
🏠
Working from home
8 followers · 16 following

Achievements

Achievement: Pull Sharkx2Achievement: Pair ExtraordinaireAchievement: YOLO

Achievements

Achievement: Pull Sharkx2Achievement: Pair ExtraordinaireAchievement: YOLO

Highlights

Block or report glatinone

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
glatinone/README.md

I'm Kiell, a cybersecurity engineer working on the security of AI agents and MCP servers: what an agent is allowed to do, what an MCP server can reach before you install it, and how to prove the boundary held. I'm based in Batam, Indonesia, and work in cybersecurity at Constellar in Singapore.

I build open source tools for this:

  • Agentic Security Lab: a deny-by-default policy boundary for agent tool calls, with an audit receipt for every decision. Try the live workbench.
  • mcpscan: a static supply-chain scanner for MCP servers and Claude Code projects. 30 rules mapped to the OWASP MCP Top 10, SARIF output, pip install mcpscan-cli.
  • mcp-auth-harness: tests an MCP server against the 2026-07-28 authorization spec. A naive baseline passes 2 of 10 checks; the hardened reference server passes all 10.
  • Agent Memory Protocol: an open protocol and reference server for shared agent memory, with per-memory access rules.

When I find a bug in a tool I use, I fix it upstream. My patches have been merged into NVIDIA SkillSpector, Tencent AI-Infra-Guard, headroom and six other projects, and I have open pull requests at NVIDIA garak and LiteLLM.

Merged upstream

headroomlabs-ai/headroom #3100 - 2026-11-06
agent0ai/a0-plugins #548 - 2026-09-16
Tencent/AI-Infra-Guard #541 - 2026-09-02
ahujasid/mcp-for-blender #324 - 2026-08-24
NVIDIA/SkillSpector #381 - 2026-08-20
zalando/skipper #4125 - 2026-07-13
gpustack/gpustack #5817 - 2026-07-10
teamhanko/hanko #2725 - 2026-07-09
kerlenton/mcpsnoop #40 - 2026-07-08

Writing

MCP Security: A Practitioner's Guide
I wrote an MCP server auth test harness for the 2026-07-28 spec
A fake MCP server spent three months earning trust
I scanned 30 Docker images. 10 still trust the root Chrome is dropping
We planted 10 vulnerabilities to test free Semgrep. It reported 3.

More at kielltampubolon.id/writing

Website · LinkedIn · DEV.to (1,052 followers) · Medium

Pinned Loading

  1. agentic-security-lab agentic-security-lab Public

    An executable, deny-by-default policy boundary for AI agent tool calls, with runtime enforcement, adversarial cases, and audit receipts.

    JavaScript 1

  2. mcpscan mcpscan Public

    Supply-chain security scanner for MCP servers and agent projects. Finds risky tools, permissions, commands, and secrets before install.

    Python

  3. mcp-auth-harness mcp-auth-harness Public

    Test harness for MCP 2026-07-28 server authorization: naive vs hardened server, 10 checks

    Python

  4. agent-memory-protocol agent-memory-protocol Public

    Shared memory for AI agents: agents forget between sessions, AMP is the open protocol (with a reference server and client SDKs) that lets them remember and share what they know. Like MCP, but for m…

    Python

  5. NVIDIA/SkillSpector NVIDIA/SkillSpector Public

    Security scanner for AI agent skills. Detect vulnerabilities, malicious patterns, security risks, prompt injection, data exfiltration, and supply-chain risks in Claude Code, Codex, and MCP skills b…

    Python 24.8k 2.2k

  6. Tencent/AI-Infra-Guard Tencent/AI-Infra-Guard Public

    A full-stack AI Red Teaming platform securing AI ecosystems via Agent Scan, Skills Scan, MCP scan, AI Infra scan and LLM jailbreak evaluation.

    Python 7.4k 741