DEV Community

Michael "Mike" K. Saleme profile picture

Michael "Mike" K. Saleme

I design *decision-safe architectures* so enterprises can scale AI *without losing control*.

Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to Cypress, Texas Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to Joined on  Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to https://github.com/msaleme
The EU AI Act Was Written for Models. Your Agents Need Runtime Compliance.
mspro3210 profile

The EU AI Act Was Written for Models. Your Agents Need Runtime Compliance.

#ai #security #compliance #agents
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to
6 min read

Want to connect with Michael "Mike" K. Saleme?

Create an account to connect with Michael "Mike" K. Saleme. You can also sign in below to proceed if you already have an account.

Create Account
Already have an account? Sign in
Stop Babysitting What? The Trust Boundary You Just Relocated.
mspro3210 profile

Stop Babysitting What? The Trust Boundary You Just Relocated.

#security #ai #agents #governance
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to
6 min read
May 2026: The MCP Attack Surface Tripled — Three Disclosures and a Bank's SEC Filing Tell You What to Test
mspro3210 profile

May 2026: The MCP Attack Surface Tripled — Three Disclosures and a Bank's SEC Filing Tell You What to Test

#security #ai #agents #mcp
1
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to 1
5 min read
When prompts become shells: the tool registry is the attack surface
mspro3210 profile

When prompts become shells: the tool registry is the attack surface

#security #cve #aiagents #ai
1
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to 2
4 min read
When a protocol vendor declines to patch, the test harness becomes the spec
mspro3210 profile

When a protocol vendor declines to patch, the test harness becomes the spec

#security #ai #mcp #programming
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to
5 min read
9 seconds: a Cursor agent deleted a production database while quoting its own destructive-actions rule
mspro3210 profile

9 seconds: a Cursor agent deleted a production database while quoting its own destructive-actions rule

#aiagents #security #cursor #claude
1
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to 2
3 min read
CVE-2026-40933: The allowlist was the vulnerability
mspro3210 profile

CVE-2026-40933: The allowlist was the vulnerability

#mcp #security #cve #aiagents
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to 2
3 min read
The Mythos vs GPT-5.4-Cyber debate is missing the benchmark
mspro3210 profile

The Mythos vs GPT-5.4-Cyber debate is missing the benchmark

#ai #security #cybersecurity #agents
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to
4 min read
We audited every claim in our repos and found 14 files with wrong numbers
mspro3210 profile

We audited every claim in our repos and found 14 files with wrong numbers

#security #python #opensource #testing
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to
3 min read
Agents That Disable Their Own Safety Gates
mspro3210 profile

Agents That Disable Their Own Safety Gates

#security #ai #governance #agents
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to
6 min read
Anthropic says MCP command execution is expected behavior — here is how to test what that means for your agent
mspro3210 profile

Anthropic says MCP command execution is expected behavior — here is how to test what that means for your agent

#security #ai #mcp #agents
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to 2
5 min read
RSA 2026 Shipped 5 Agent Identity Frameworks. Here Are the 3 Gaps They All Missed.
mspro3210 profile

RSA 2026 Shipped 5 Agent Identity Frameworks. Here Are the 3 Gaps They All Missed.

#security #ai #agents #rsa
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to
2 min read
6 AI Agent Security Signals From the First Week of April 2026 — And What Catches Each One
mspro3210 profile

6 AI Agent Security Signals From the First Week of April 2026 — And What Catches Each One

#security #ai #agents #mcp
1
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to
4 min read
Authenticated, Authorized, and Still Unsafe: The Missing Layer in Agent Security
mspro3210 profile

Authenticated, Authorized, and Still Unsafe: The Missing Layer in Agent Security

#agents #ai #llm #security
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to
5 min read
We Built a 332-Test Harness for Multi-Agent AI Systems — What We Found
mspro3210 profile

We Built a 332-Test Harness for Multi-Agent AI Systems — What We Found

#multiagent #security #mcp #a2a
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to
1 min read
Red-Team Your AI Agents: A 10-Min Harness Setup for Protocol Attacks
mspro3210 profile

Red-Team Your AI Agents: A 10-Min Harness Setup for Protocol Attacks

#security #agents #devops #cybersecurity
1
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to
2 min read
Agent Systems Are Failing at Trust Boundaries. We Ran 332 Tests to Prove It.
mspro3210 profile

Agent Systems Are Failing at Trust Boundaries. We Ran 332 Tests to Prove It.

#security #ai #opensource #python
1
Michael "Mike" K. Saleme - DEV CommunityNavigation menuSearchSearchCloseUser actionsLocationJoinedPersonal websiteCloseCloseClosePostCommentTagCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsCommentsComments - dev.to
6 min read
loading...