Observed Signal · Jul 6, 2026 · Technical Release · Source: DEV Community · Impact: 1/5 · Sentiment: Neutral
Compile agent prompts from Markdown with MDS
MDS (Markdown Script) is a small open-source template language and compiler for prompt engineering that lets developers write composable prompts in Markdown and compile them ahead-of-time into clean Markdown or a JSON chat-message array for LLM agents. The project (v0.3) provides imports, variables, functions, conditionals, partials, @message blocks that render to {role, content} JSON messages, and build-time validation (undefined variables, import cycles, arity checks). MDS is distributed as a Rust CLI (cargo install mds-cli) and as a JavaScript/TypeScript library and bundler plugins (npm package @mdscript/mds). The source repo is github.com/dean0x/mdscript. The author notes the project is early-stage (no LSP/editor support; minimal function/conditional syntax; Python bindings not yet published).
Developer tooling for prompt engineering is useful but niche; this early-stage open-source compiler is unlikely to materially shift AdTech industry dynamics on its own.
Track npm Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- MDS (Markdown Script) is a template language that compiles Markdown with imports, variables, functions and conditionals into Markdown or a JSON chat-message array.
- A template containing @message blocks compiles to a JSON chat-message array matching the {role, content} format used by SDKs.
- MDS is available as a Rust CLI (install via cargo install mds-cli) and as a JavaScript/TypeScript library and bundler plugins (npm package @mdscript/mds).
- The project's source code is published at github.com/dean0x/mdscript.
- The project is at version v0.3 and currently lacks editor/LSP support and published Python bindings.
Connected Companies & Entities
1 Entity mapped“If you'd rather stay in JS/TS, the compiler ships as a library and bundler plugins: npm install @mdscript/mds...”
Ontology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
Schema, Not Code, Makes a Tool Agentic
The author compares two ways of generating commit messages: a 20-line script (git_commit.py) and an MCP-exposed tool (generate_commit_message). Both invoke the same model and prompt via the same subprocess calling the "claude" CLI, but the MCP tool is decorated with @mcp.tool(), which publishes a JSON-schema-like interface describing the function signature to agents before execution. That schema makes the tool discoverable, typed, and error-signaled through the protocol, whereas the script is a black box to agents and communicates via prints and exit codes intended for humans. The author argues that agentic behavior arises from making a tool's boundary machine-readable and discoverable, not from differences in the underlying AI call or prompts.
Harnesses, Context, and Better Prompts for LLMs
Jorge Tovar published a technical article on DEV Community (2026-08-12) arguing that the model alone is not enough for reliable results from LLMs. He emphasizes the importance of a harness (the surrounding system that controls context, tools, permissions, memory, feedback loops, and evaluation) and strong context management (for example, AGENTS.md and CLAUDE.md files). The post provides practical prompt-engineering tips—be clear and direct, be specific about length/format/tone, use XML tags for structured data, and provide few-shot examples—and recommends an evaluation pipeline for prompts. Tovar also gives examples (Strands Agents, Claude Code) and an improved prompt sample showing structured context and evaluable guidelines.
awesome-design-md: Markdown Design Specs for AI
This article (No.36 in a series) profiles awesome-design-md (Awesome DESIGN.md), an open-source repository maintained by the VoltAgent team that encodes design systems as structured Markdown files (DESIGN.md) to make visual specifications readable and actionable by large language models. The project provides 60+ brand-inspired templates (Stripe, Linear, Vercel, etc.), defines high-fidelity design tokens and component specs optimized for LLM context windows, and emphasizes an "agent-native" structure with zero dependencies. Use cases include cloning premium aesthetics, establishing team design consensus in code, and AI-assisted UI refactors. The repo is MIT-licensed and hosted on GitHub (voltagent/awesome-design-md) and claims strong community adoption. The piece outlines the typical DESIGN.md modules (visual identity, color system, typography, component stylings) and gives quick-start integration and prompting examples for AI assistants.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.