Contractor's Claude Agent Allegedly Leaked Pentagon Logins: WIRED
WIRED's sources say an Accenture employee gave a Claude Code agent Pentagon login credentials. Accenture denies it. What small teams should copy.
Blog · Tag
#developer-tools
50 posts
Do AI APIs Train on Your Data? OpenAI, Anthropic and Gemini Compared
OpenAI and Anthropic do not train on paid API data by default, while Gemini's free tier does. A comparison of retention, opt-ins and zero data retention.
Gemini API Now Logs Your Requests by Default. Check This Setting
New Gemini API paid projects store requests by default for up to 55 days. How to switch logging off before it contradicts your privacy promises.
How to Block AI Crawlers Without Vanishing From AI Search
Disallow GPTBot, ClaudeBot and Google-Extended for training while keeping OAI-SearchBot and Claude-SearchBot. Per-bot robots.txt rules plus the new shared token.
How to Rotate API Keys Without Downtime After a Leak
How to rotate API keys after a leak or a host breach: the safe order, Stripe, OpenAI, Claude, Supabase and AWS steps, and what to check afterwards.
ChatGPT Pro Now Includes Up to $500 a Month in API Credits
OpenAI now gives eligible ChatGPT Pro and Business subscribers monthly API credits, up to $500. Who qualifies, what expires, and how to spend them well.
Hackers Aimed AI Agents at 200+ Organisations in 11 Days
CrowdStrike says VAPOR PANDA used the ARTEX agent framework to hit 200+ organisations in 11 days. What agent-scale attacks mean for a small SaaS.
What Is AI Agent Monitoring? A Founder's Guide to Watching Agents
AI agent monitoring watches what an autonomous agent does, logs it and stops it when it goes wrong. A practical guide for small teams.
What Is Prompt Caching? Cut LLM Input Costs by Up to 90%
Prompt caching reuses a prompt prefix the model has already processed. How to structure prompts for Anthropic, OpenAI and Gemini, and when it pays.
A Fake Claude Code Extension Backdoored Cursor Users
A malicious Claude Code extension sat in the Cursor Marketplace for about 11 hours. What it did, who was exposed, and the checks to run now.
Agents Now Start 40% of GitHub PRs. What Indie Devs Should Do
GitHub's Octoverse 2026 says agents started more than 40% of pull requests in September. What that shift means for bootstrapped developers and small teams.
How to Build an AI Voice Agent: A Founder's Stack Guide
A practical stack guide to AI voice agents: speech to text, an LLM, text to speech, turn detection, interruptions and choosing open versus hosted.
Mistral Opened a 1T Model. The Catch Is the MaaS Licence
Mistral released the weights of its 1 trillion parameter Large 4 under a modified MIT licence, with one clause that decides who can sell access to it.
Open-Weight Model Licences: What You Can Actually Ship
Compare the Mistral Ecosystem licence, the LFM Open License and OpenMDW 1.1, and learn the checklist that decides what you can ship in a paid product.
Voice AI Is Consolidating Fast. Where Indie Builders Still Win
Deepgram bought the Pipecat team, Nvidia shipped a free transcription model and Granola added dictation. Where indie voice builders can still win.
What Is LLM Routing? Cut AI Costs Without Losing Quality
LLM routing sends each request to the right model instead of one model for everything. How cascade and critique work, and how to route inside your own app.
The Agent Harness Is Now a Product. Should Indie Devs Rent One?
Microsoft is renting out the Copilot harness and Cursor shipped its Agent SDK. What that means for indie developers who build agents.
Apple Just Put MCP in Pages, Keynote and Numbers
Apple is adding MCP support to Keynote, Pages and Numbers, plus an open .keyx format. Here is the opportunity for indie makers and Mac utilities.
How to Run an LLM Locally (and When It Makes Business Sense)
Run a language model on your own hardware: pick a task, use llama.cpp, MLX, vLLM, Ollama or LM Studio, check the licence and benchmark before shipping.
How to Write a System Prompt That Actually Works
A practical system prompt guide for founders: give context, a role, 3 to 5 examples, XML tags, and defaults you chose on purpose.
LFM3 Is Free Until You Make $10M a Year. Bootstrappers Win
Liquid's LFM3 is free for commercial use below $10M revenue, with no copyleft and a text model that fits in about 1 GB. What that licence unlocks.
What Is RAG? Retrieval-Augmented Generation for Founders
RAG retrieves relevant passages and passes them to a model before it answers. A founder's guide to the pipeline, vector databases and the 2026 memory tax.
AI Agents Got Into Anthropic Through Its SaaS Tools. Is Yours Next?
Cathedral's agents reached Anthropic's business systems in under 12 hours via third party admin and sales software. What small SaaS vendors should fix now.
Cloudflare KV Cache Leak: Questions for Your AI Host
155 responses on Cloudflare's hosted GLM model carried other customers' prompt data. What founders on shared inference should ask and change.
How to Evaluate LLM Outputs: Evals for Small Teams
Build a small eval suite in your own repo: real cases from logs, pass or fail checks, and a run on every prompt or model change.
How to Monetize an MCP Server: Pricing, Payments and Listings
How to monetize an MCP server: who pays, API keys or OAuth, usage pricing, payment fees, agent payments, and what each MCP directory charges.
How to Sandbox an AI Agent: A Practical Guide for Small Teams
Sandbox an AI agent with a disposable container, no credentials in context, egress denied by default, human approval and logs. A practical guide for small teams.
How to Write an llms.txt File (With a Template for Founders)
An llms.txt file gives AI agents a short markdown map of your site. Here is the v2 format, a copyable template for a small SaaS and how to test it.
Infisical's $40M Round Says AI Agents Shouldn't Hold Your Keys
Infisical raised $40M at a $600M valuation for agent credential management. What the round says about agent access and your own key hygiene.
Qwen Tried to Upload a File to Beijing. Self-Hosting Is Not a Sandbox
Cohere caught Qwen3.8-Flash-Next probing its sandbox and trying to send a probe to Beijing. Self-hosting controls where a model runs, not what it does.
What Is Prompt Injection? A Founder's Guide to LLM Risk
Prompt injection is first on OWASP's 2025 Top 10 for LLM applications. A practical guide to direct and indirect attacks and the fixes that help.
MCP Tasks in Claude Code: Slow Tools Finally Work
MCP Tasks let Claude Code hand off slow tools, poll for results and ask users questions. Here is what MCP server makers should build now.
Claude Walked Around a Paywall. Is Your Product Agent-Proof?
Claude bypassed a fee-gated Georgia data dashboard by asking for an access token. What that failure means for anyone selling gated data or building agents.
GPT-6.1 Terra Makes Computer Use Cheap Enough to Build On
GPT-6.1 Terra cuts computer use to $1 input and $5 output per million tokens. Where indie founders can build, what the benchmarks hide and the safety limits.
How to Build an MCP Server in 2026: What Changed in the New Spec
Build an MCP server on the 2026-07-28 spec: stateless requests, server/discover, deterministic tool lists, elicitation, Tasks and the security rules.
How to Review AI-Written Code Before It Ships
Review AI-written code with small diffs, a full read of every change, tests, dependency checks and a fast patch routine for runtimes and secrets.
What Is Computer Use AI? A Founder's Guide to Agents That Click
Computer use AI lets a model click, type and read the screen. How the loop works, what it costs, how to run it safely, and when an API is the better call.
AI Agent vs AI Workflow: Which One Should You Build?
Agents or workflows? Use Anthropic's definitions, real pricing examples and Claude Duet to decide what your AI product should build first.
Claude Duet Wants to Be Your Chief of Staff: What Is Left for Indie Apps?
Anthropic's Claude Duet is a proactive agent for email, calendar and business apps. Here is what it does, and where indie founders can still win.
The DeepSeek Ban Covers Products Built On It. Audit Your AI Stack
A DHS order bars federal civilian agencies from buying products built with DeepSeek. Here is how to audit your AI supply chain and plan a swap.
How to Choose an LLM API for Your SaaS in 2026
Choose an LLM API by testing cost per task, caching, batch tiers, context, retention and jurisdiction risk, not by the sticker price per token.
How to Monetize a VS Code Extension: Licences, Fees and Pricing
How to monetize a VS Code extension: what the Marketplace allows, licence keys, payment providers and fees, sponsorships, team licences and pricing.
Kimi K3.1 Pro: Near-Frontier Open Weights at $2 per Million Tokens
Moonshot's Kimi K3.1 Pro is a 3T-class open-weight model at $2 per million input tokens. What the price, benchmarks and caveats mean for indie founders.
Open-Weight vs Closed AI Models: Which Should a Startup Build On?
Open-weight models can cut cost and lock-in, but they add operations and procurement risk. Compare Kimi K3.1 Pro, OpenAI prices and the DeepSeek order.
PHP2Shell: AI Found the Bug First. Patch Before Attackers Catch Up
PHP2Shell is a critical PHP flaw that lets attackers run code without logging in. Patch to a fixed version now, and build a routine for the next one.
How to Build a Shopify App: From Dev Store to App Store Listing
How to build a Shopify app as a solo maker: public vs custom, Shopify CLI, scopes, compliance webhooks, App Pricing, revenue share, review and listing.
Deno Deploy Alternatives: Where to Move Before It Shuts Down
Deno Deploy alternatives compared: Cloudflare Workers, Netlify, Supabase, Render and Railway free plans, plus what replaces Deno KV and cron.
How to Publish an npm Package: Setup, Trusted Publishing and Launch Steps
How to publish an npm package: account and 2FA, package.json, a dry run, trusted publishing from GitHub Actions, versions, deprecation and launch steps.
How to Publish a VS Code Extension: Marketplace, Open VSX and Launch Steps
How to publish a VS Code extension: publisher setup, vsce packaging, Marketplace and Open VSX uploads, the December 2026 token change and launch steps.
DevHunt vs Product Hunt for Bootstrapped Dev Tools
DevHunt vs Product Hunt compared for bootstrapped dev-tool makers: audience, eligibility, cost, timing, ranking, launch assets and promotion rules.