sunnydachs

I enjoy building things for fun and exploring new technologies.

Japan Joined on Sep 5, 2026 https://github.com/sunnydachs

sunnydachs

Nov 3

Do AI-Written Tests Actually Catch Bugs? I Measured It

#testing #python #ai #mutationtesting

4 min read

Want to connect with sunnydachs?

Create an account to connect with sunnydachs. You can also sign in below to proceed if you already have an account.

Create Account

Already have an account? Sign in

sunnydachs

Nov 1

Cut two things out of an agent loop and it stops working

#agents #ai #llm #testing

5 min read

sunnydachs

Oct 28

Don't trust the agent's "done": receipt checks, two models

#agents #ai #testing #programming

6 min read

sunnydachs

Oct 25

Your Agent Says Done. The Cron Tab Disagrees.

#agents #ai #testing #programming

5 min read

sunnydachs

Oct 20

A leak notice arrived. What to do today, this week, and in 30 days

#cybersecurity #infosec #privacy #security

4 min read

sunnydachs

Oct 19

Only one of four gates stopped every attack on an AI agent

#agents #ai #llm #security

4 min read

sunnydachs

Oct 14

Who's Auditing the Skills Your AI Agent Runs?

#agents #ai #security #opensource

5 min read

sunnydachs

Oct 6

I Moved Decision Logic Out of an LLM and Measured What Broke

#ai #machinelearning #testing

5 min read

sunnydachs

Oct 6

Your Agent Says Verified. The Check Behind It May Be Empty.

#agents #ai #security #testing

5 min read

sunnydachs

Oct 5

Your agent says verified. Nothing binds it to the artifact

#agents #ai #security #testing

6 min read

sunnydachs

Oct 3

span-01 vs mercury-decide: same score, opposite failures

#ai #machinelearning #testing

4 min read

sunnydachs

Oct 3

Your tool returned the rows. The model counted them wrong.

#agents #ai #llm #software

6 min read

sunnydachs

Sep 30

Your agent says "done" — and nothing ran

#agents #ai #debugging #software

5 min read

sunnydachs

Sep 28

Is regex enough? I tested span-01 on mixed-language text

#ai #machinelearning #programming

6 min read

sunnydachs

Sep 25

I changed a tool's schema four ways and watched three agent frameworks break differently — 36 runs measured

#aiagents #llm #langgraph #python

5 min read

sunnydachs

Sep 25

AI answers are easy to get but expensive to read — a CLI that filters them down

#ai #cli #productivity #tools

3 min read

sunnydachs

Sep 23

Do agents survive a crash, and why does an LLM retry execute the same side effect twice? 34 runs measured

#aiagents #llm #langgraph #python

4 min read

Real-world benchmarks reveal hidden framework bugs

sunnydachs

Sep 21

What happens when enterprise requirements hit Strands, LangGraph, and CrewAI - 45 runs measured

#agents #ai #llm #python

4 min read

sunnydachs

Sep 21

I built the same agent in Strands, LangGraph, and CrewAI — and recorded every LLM call to see how they actually differ

#aiagents #langgraph #crewai #python

4 min read

sunnydachs

Sep 20

Your GDScript arrays are slower than you think? — I built a CLI that detects the pitfalls and benchmarks them

#cli #gamedev #performance #tools

4 min read

sunnydachs

Sep 19

Your OpenAPI spec is a contract. Nobody is checking it. I built a CLI that detects spec↔code drift with zero dependencies.

#python #api #openapi #opensource

4 min read

sunnydachs

Sep 18

The tracking plan said "implemented". The dashboard says otherwise. I built a CLI that detects plan↔code drift using only AST.

#python #cli #analytics #opensource

3 min read

sunnydachs

Sep 16

Your README code examples are silently lying — I built a CLI to detect documentation drift using only AST, no LLM

#cli #documentation #softwaredevelopment #tooling

3 min read

sunnydachs

Sep 13

Audit your Excel pivot tables before they break — a read-only CLI with zero dependencies

#excel #python #opensource #devtools

3 min read

sunnydachs

Sep 5

I automated Dependabot PR cleanup — and drew a hard line on what NOT to automate

#dependabot #github #devops #python

4 min read

DEV Community

sunnydachs

Badges

4 Week Community Wellness Streak

2 Week Community Wellness Streak

1 Week Community Wellness Streak

Writing Debut

Skills/Languages

Currently learning

Currently hacking on

Available for

Do AI-Written Tests Actually Catch Bugs? I Measured It

Want to connect with sunnydachs?

Cut two things out of an agent loop and it stops working

Don't trust the agent's "done": receipt checks, two models

Your Agent Says Done. The Cron Tab Disagrees.

A leak notice arrived. What to do today, this week, and in 30 days

Only one of four gates stopped every attack on an AI agent

Who's Auditing the Skills Your AI Agent Runs?

I Moved Decision Logic Out of an LLM and Measured What Broke

Your Agent Says Verified. The Check Behind It May Be Empty.

Your agent says verified. Nothing binds it to the artifact

span-01 vs mercury-decide: same score, opposite failures

Your tool returned the rows. The model counted them wrong.

Your agent says "done" — and nothing ran

Is regex enough? I tested span-01 on mixed-language text

I changed a tool's schema four ways and watched three agent frameworks break differently — 36 runs measured

AI answers are easy to get but expensive to read — a CLI that filters them down

Do agents survive a crash, and why does an LLM retry execute the same side effect twice? 34 runs measured

What happens when enterprise requirements hit Strands, LangGraph, and CrewAI - 45 runs measured

I built the same agent in Strands, LangGraph, and CrewAI — and recorded every LLM call to see how they actually differ

Your GDScript arrays are slower than you think? — I built a CLI that detects the pitfalls and benchmarks them

Your OpenAPI spec is a contract. Nobody is checking it. I built a CLI that detects spec↔code drift with zero dependencies.

The tracking plan said "implemented". The dashboard says otherwise. I built a CLI that detects plan↔code drift using only AST.

Your README code examples are silently lying — I built a CLI to detect documentation drift using only AST, no LLM

Audit your Excel pivot tables before they break — a read-only CLI with zero dependencies

I automated Dependabot PR cleanup — and drew a hard line on what NOT to automate