SafeLine WAF Can Be Added to Any Docker Compose Stack in Three Steps

·7 views

Developers running existing Docker Compose stacks can integrate the SafeLine Web Application Firewall without rebuilding or modifying their app containers. The process involves connecting existing services to a shared Docker network, installing SafeLine via its official script, and linking its reverse proxy to that shared network. Nginx or other reverse proxies are then configured to route traffic through SafeLine before it reaches the application, creating an inspection layer between the internet and the app. SafeLine operates as an independent Compose project, meaning updates to either the app stack or the WAF do not interfere with each other. The setup also supports a detection-only mode, allowing teams to monitor for false positives before enabling active traffic blocking.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Open-source tool tokenscope helps developers track hidden AI agent context costs

A new open-source CLI tool called tokenscope lets developers scan their AI agent codebases and estimate the token footprint of prompts, tool schemas, and configuration files. The tool addresses a growing cost problem: most AI agent frameworks resend accumulated context on every model call, meaning a single bloated prompt multiplies across every request and every user. Developers can run tokenscope locally with a single npx command, integrate it as a pre-push Git hook, or deploy it as a GitHub Actions check to flag token-cost increases on pull requests. The tool is tokenizer-free and provides relative estimates rather than exact billing figures, requiring no API keys or provider SDK setup. It is released under the Apache-2.0 license and is available on npm and GitHub.

0 comments Read more at DEV Community

ProgrammingHacker News ·

France bans unsolicited telemarketing calls starting August 11

France is set to implement a ban on unsolicited telemarketing calls, with the measure taking effect on August 11, 2026. The move aims to protect consumers from unwanted commercial solicitations by phone. The policy represents a significant shift in how businesses are permitted to reach potential customers in France. The ban reflects growing public frustration with intrusive telemarketing practices across the country.

0 comments Read more at Hacker News

ProgrammingDEV Community ·

AI Evals Explained: Why Product Teams, Not Engineers, Must Define Quality

AI evaluations (evals) are structured systems that track whether an AI product's output quality changes over time, helping teams avoid subjective debates about performance degradation. At their core, evals consist of real user queries, written criteria for what a good response looks like, and a repeatable method to check outputs against those criteria. The central challenge is not technical but organizational: someone must explicitly define what 'correct' means for each use case, including tone, accuracy, and acceptable trade-offs. Experts argue that product managers should own these quality definitions in plain prose, while engineers handle the automated testing infrastructure. Teams that fail to assign clear ownership risk having product judgements made by default, embedded silently into dashboards and test scripts by whoever built them.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Kimi K3's 2.81T Parameters Put It Far Beyond Any Mac's Memory Capacity

Kimi K3 is a mixture-of-experts model with 2.81 trillion parameters, requiring roughly 1.4 TB of memory even at 4-bit quantization — far exceeding the 512 GB maximum available on Apple's most powerful Mac Studio. Despite MoE architecture activating only a small fraction of experts per token, all experts must remain in memory since the router can call any of them at any moment, meaning sparse activation reduces compute but not memory footprint. Ollama's only available Kimi K3 tag is labeled ':cloud', which silently routes requests to Moonshot's remote servers rather than running inference locally on the user's machine. In practical testing on a 128 GB MacBook M4 Max, the Qwen2.5-Coder 14B model delivered 13.3 tokens per second versus 5.6 tok/s for the 32B variant, making the smaller model significantly more usable for agentic workflows. The 14B model also fits within 16 GB of unified memory, meaning the base $599 Mac mini can run a capable local coding model despite being unable to handle frontier-scale models like Kimi K3.

0 comments Read more at DEV Community