claude.ai

Open-source compiler for AI agent communication

Developer Ibrahim Kobeissy published /calibrate-comms, an open-source skill implemented inside an Obsidian vault that captures and persists user preferences for how AI agents should write. The tool calibrates nine operational dials (Density, Sequence, Modality, Abstraction, Tradeoff, Detail, Jargon, Tone, Context-giving) via a Quick mode (forced-choice proxies) or Deep mode (fetching open-access instruments when available). Pairwise sample reactions override priors, and a deterministic band-to-rule compiler emits executable directives written into a delimited COMMS-PROFILE block (stored in CLAUDE.md) so multiple agents (notably Claude Code and Codex) can share one source of truth. The author emphasises privacy by keeping populated profiles in a private vault while shipping the public template and an empty placeholder on GitHub. Publication date: 2026-06-26.

AI Models Still Recommend Twilio for Notifications

Bersyn tested four AI models (ChatGPT, Claude, Gemini, Perplexity) with developer-style prompts asking for the best multi-channel notification API for a B2B SaaS platform. In the experiment (published June 26, 2026), three models mainly recommended incumbents such as Twilio and OneSignal, while Anthropic’s Claude was the outlier and named modern notification providers (Knock, Novu, Courier) far more often. Bersyn highlights a table of model responses showing Claude frequently naming Knock (80%), Novu (80%) and Courier (60%), whereas ChatGPT omitted Knock and Courier entirely and barely mentioned Novu. The author argues this discrepancy matters because buyers who ask LLMs may never learn about newer category-specific tools, so companies must create verifiable, third-party content that ties them to the category to appear in model recommendations.

Old Notes Gain New Meaning via AI

A dev.to author reflects on how connecting their long-running personal notes to an AI changed the usefulness of that archived material. The writer has used note-taking tools for years (Evernote, Notion, Logseq, Taskuma) and recently connected Notion and Google Docs via a tool called MCP. Allowing the AI model Claude to read accumulated reviews, task notes and brainstorms made past entries actionable: Claude now draws on earlier reasoning to answer current questions, ground decisions, and improve brainstorming. The author describes this as old content acquiring new value once a reader (the AI) existed, easing anxieties that prior work might have been wasted. Publication date recorded as 2026-06-26.