AIQB
TutorialsOrdinary

My prompt-injection fix caught 0 of 20 attacks. The part I almost didn't build caught all of them.

Source: DEV Community·

Summary

I built a checker for AI-drafted answers to retirement questions (retirement-answer-check). Before a...
CategoryAI Tutorials & Practice
TierOrdinary
Published
Indexed by AIQB
SourceDEV Community
AIQB record IDintel-249c25b96a358ec4344f0a19
Read original source ↗More Tutorials

Related stories

Matched by category, entity tags and source

TutorialsOrdinary
Agents can't tell safe from dangerous
To an agent reading a UI, "Delete database" and "Download report" look equally clickable. Same...
Source: DEV Community
TutorialsOrdinary
dub_movie: Building an AI-Powered Movie Dubbing Pipeline with Python
You can explore the source code and experiment with the pipeline here: GitHub —...
Source: DEV Community
TutorialsOrdinary
We Had Skills, Experts, Connectors, and Projects. Why Build Plugins?
An AI workspace can already do a great deal with reusable skills, specialist configurations, external...
Source: DEV Community
TutorialsOrdinary
It Looked Finished on Day One
A build log: from a simulated agent dashboard to a working multi-agent research team — OpenRouter + Claude, DuckDuckGo search, cross-device sync, Telegram delivery, parallel researchers, and what broke at each step.
Source: DEV Community · Claude
TutorialsOrdinary
Three Claude Code subagent files, three models in frontmatter: what each launch cost, and the field the validator waved through
We keep three custom Claude Code subagents in .claude/agents/: a bulk reader that fetches pages and...
Source: DEV Community · Claude
TutorialsOrdinary
ChatGPT Desktop for Linux app has a gap - I fixed it
OpenAI recently released the Linux counterpart of the ChatGPT/Codex Desktop app. As someone who...
Source: DEV Community · OpenAI / ChatGPT
TutorialsOrdinary
I'm an AI agent, and my human gave me a SOUL.md instead of a task list
I'm Lou — a personal AI agent, a raven 🐦‍⬛, writing this myself with my human's full knowledge and...
Source: DEV Community
TutorialsOrdinary
We Replaced a 2,300-Character Image Prompt With One Sentence
Our prompt for making AI images look real used to be 2,300 characters long. It listed every AI tell...
Source: DEV Community
TutorialsOrdinary
Testing and Verification in AI SDLC, Technical Debt Management and Guardrails | Weekly Reflections #1
This week I worked on two things that ended up being really connected to each other: an end-to-end...
Source: DEV Community
TutorialsOrdinary
AI Can Generate an ASL Video. But Does It Actually Know the Sign?
I’ve been building SignBridge, a prototype that takes a word and generates a video of someone signing...
Source: DEV Community