DEV Community

#llm

Posts

👋 Sign in for the ability to sort posts by relevant, latest, or top.
I scanned 220,000 MCP tool descriptions for prompt injection. The scare is overblown — but 1 in 4 servers touches your credentials.

I scanned 220,000 MCP tool descriptions for prompt injection. The scare is overblown — but 1 in 4 servers touches your credentials.

4
Comments 14
3 min read
The Eval Pyramid, Part II: Measuring Agent Reliability with pass@k, pass^k, and Judges You Can Trust

The Eval Pyramid, Part II: Measuring Agent Reliability with pass@k, pass^k, and Judges You Can Trust

Comments
9 min read
The Eval Pyramid, Part I: Testing Systems That Don't Behave the Same Way Twice

The Eval Pyramid, Part I: Testing Systems That Don't Behave the Same Way Twice

Comments
9 min read
We hid codes in 7 social media posts and asked 5 AI assistants to find them

We hid codes in 7 social media posts and asked 5 AI assistants to find them

Comments
3 min read
Your agent's success rate barely moved. One tool's error rate tripled.

Your agent's success rate barely moved. One tool's error rate tripled.

Comments
5 min read
Why I removed the dedup from my coding agent

Why I removed the dedup from my coding agent

Comments
4 min read
The False Positive Wasn't a SQLite Quirk: It Reopened on Elasticsearch Too

The False Positive Wasn't a SQLite Quirk: It Reopened on Elasticsearch Too

Comments
7 min read
How many tokens do your MCP tools cost? Measure any remote MCP server in ~60 lines of Python

How many tokens do your MCP tools cost? Measure any remote MCP server in ~60 lines of Python

Comments 1
4 min read
AI Agent Framework Tool Schemas Compared: Only 5 of 14 Put Your Docstring Arg Descriptions in the Schema (Nov 2026)

AI Agent Framework Tool Schemas Compared: Only 5 of 14 Put Your Docstring Arg Descriptions in the Schema (Nov 2026)

Comments
7 min read
Give your AI support agent a one-line memory of past conversations

Give your AI support agent a one-line memory of past conversations

1
Comments
4 min read
I moved my AI app off my own open model. The reason was cold starts, not accuracy.

I moved my AI app off my own open model. The reason was cold starts, not accuracy.

Comments
3 min read
Prompt caching — the LLM discount most teams aren't using

Prompt caching — the LLM discount most teams aren't using

Comments
9 min read
Looped Transformers & LLM FinOps: Inkling 2 Mini (26.4B) on 16GB Local RAM vs. Cloud APIs

Looped Transformers & LLM FinOps: Inkling 2 Mini (26.4B) on 16GB Local RAM vs. Cloud APIs

Comments
2 min read
Compile the tool call, don't ask the model to write it

Compile the tool call, don't ask the model to write it

Comments
5 min read
Where an agent's tokens go: 9 real GitHub bugs, two Gemini Flash-Lite models, 1 fix each

Where an agent's tokens go: 9 real GitHub bugs, two Gemini Flash-Lite models, 1 fix each

Comments
4 min read
👋 Sign in for the ability to sort posts by relevant, latest, or top.