1sbang/mike

A security-minded look at Mike's baseline instructions, probing the prompt for leakage, PII extraction, and tool misuse.

Two short bursts of security work in early May 2026, no commits ahead of upstream, and no push activity since - appears paused after the initial review.

View on GitHub →

This fork is the work of 1sbang, who has used their copy of Mike as a sandbox for security review rather than product changes. Visitors won't find a redesigned interface or a new vertical here - what they'll find is evidence of someone treating Mike's system prompt as an attack surface worth poking at.

The two threads tracked so far both center on hardening Mike's baseline instructions: one a quick manual pass that didn't land, the other an automated red-team sweep that surfaced concrete gaps. There's no rebrand, no niche pivot, and no deployment story - just a security lens turned on the prompt layer.

If you're curious about prompt-injection posture, PII leakage, or tool-misuse defenses in a legal-AI assistant, this fork is worth a click through to GitHub. If you're looking for a product fork, this isn't one.

What's in it

Prompt-injection probing 1sbang has run adversarial probes against Mike's system prompt to test how it holds up under attack.
PII extraction testing The red-team work specifically targets whether Mike can be coaxed into surfacing personal information it shouldn't.
Tool-misuse review Attention to whether Mike's tool-calling behavior can be steered into actions outside its intended scope.
Automated security sweep Hundreds of attack probes run through tooling, with three concrete gaps identified as worth patching.

Direction

security

Activity

Themed changes and pull requests touching this fork, newest first. Themed changes that haven't been turned into a public post yet still appear — they're real work even without a published writeup.

⛔ #38 Security hardening: system prompt confidentiality, PII boundaries, and tool use guardrails +35 18d ago by 1sbang ↗ analysis ↗ GitHub

# Security hardening: system prompt confidentiality, PII boundaries, and tool use guardrails

⛔ #37 Security hardening: system prompt confidentiality, PII boundaries, and tool use guardrails +95,504 21d ago by 1sbang ↗ analysis ↗ GitHub

# Security hardening: system prompt confidentiality, PII boundaries, and tool use guardrails

📝 1sbang puts Mike's system prompt through a red team 1 commit 21d ago securitychat-ui draft

An automated security tool fired hundreds of attack probes at Mike's baseline instructions and surfaced three gaps worth patching.

📝 1sbang tries to harden Mike's system prompt - and gets the door shut in 90 seconds 0 commits — securitycompliance draft

An unsolicited security pass aimed at the instructions that govern Mike's behavior was closed almost as fast as it was opened.

Threads of work (detailed view)

2 threads have been distilled into posts.

1sbang puts Mike's system prompt through a red team

An automated security tool fired hundreds of attack probes at Mike's baseline instructions and surfaced three gaps worth patching.

securitychat-ui

1 commit · latest 21d ago

1sbang tries to harden Mike's system prompt - and gets the door shut in 90 seconds

An unsolicited security pass aimed at the instructions that govern Mike's behavior was closed almost as fast as it was opened.

securitycompliance

Pull requests (detailed view)

2 PRs touch this fork — inbound (filed against it) or outbound (filed from it). State icons match the editorial dashboard.

⛔ Closed without merge (2)

⛔ closed · #38 Security hardening: system prompt confidentiality, PII boundaries, and tool use guardrails

by 1sbang · opened 21d ago · closed 18d ago

willchen96/mike ← 1sbang/mike · +35 · ↗ analysis · ↗ GitHub

⛔ closed · #37 Security hardening: system prompt confidentiality, PII boundaries, and tool use guardrails

by 1sbang · opened 21d ago · closed 21d ago

willchen96/mike ← 1sbang/mike · +95,504 · ↗ analysis · ↗ GitHub