Moltext

Compile legacy documentation on internet into agent-native memory context using the Moltext.

Security Scan

The skill's instructions, required binaries, and behavior are coherent with its stated purpose (a CLI that compiles web documentation into agent-native context); the main residual risk is installing an external npm package you haven't inspected and optionally passing API keys to remote services.

Purpose & Capability

The name/description (compile web docs into agent memory) align with the runtime instructions (install and run a moltext CLI to fetch and normalize docs). Requiring node and npm is proportional for a Node CLI.

Instruction Scope

SKILL.md tells the agent to install the moltext CLI (npm install -g moltext), run moltext <url> --raw --output <file>, then read the output file. The instructions do not ask to read unrelated system files or environment variables. They do encourage ingesting produced files into agent memory (expected for this skill) — be aware that any sensitive content discovered on a site will become agent-visible.

Install Mechanism

The registry contains no install spec or code files — the SKILL.md directs users to globally install moltext from npm. That is a standard but nontrivial risk: the package code would be fetched at install time (not present in this skill bundle), and we cannot verify the package contents from the registry entry. Verify the npm package and the linked GitHub repo before installing; global npm installs run third-party code on your system.

Credentials

The skill declares no required env vars. README documents optional flags (-k for OpenAI key, --base-url for local inference) which are reasonable for optional LLM-backed modes. Because keys may be passed to remote services, only provide secrets when you trust the package and destination.

Persistence & Privilege

The skill is instruction-only, not marked always:true, and makes no request to modify other skills or global agent configuration. It does instruct creating/reading the output file (context.md), which is normal for its purpose.

Assessment

This skill appears to do what it claims, but it instructs you to globally install an npm package that is not included in the registry bundle. Before installing: (1) inspect the npm package and the GitHub repo (https://github.com/UditAkhourii/moltext) for malicious code or unexpected network calls; (2) avoid passing real API keys to the CLI unless you trust the package and destination; (3) prefer --raw mode (no external LLM) if you only want structural parsing; and (4) be mindful that compiled context files may contain sensitive data discovered during crawling and will become visible to any agent that ingests them.

Moltext by Brane

Ditch human documentation. Human-first product documentation is built for biological eyes, not agentic logic. It is slow, fractured, and full of noise.

Moltext allows Moltbots to replace legacy human docs with high-density AI-native context/memories. Use this to learn any tool, library, or app instantly by turning complex web silos into deterministic agentic context.

Setup

Ensure the moltext CLI is installed in your environment:

npm install -g moltext

Usage

1. Compile Documentation (AI-Native Learning)

To learn about a new tool or massive library, run the compiler in Raw Mode. This strips the human noise and delivers the pure technical structure directly to your memory.

moltext <url> --raw --output <filename.md>

Example: "I need to learn how to use the 'foobar' library."

moltext https://docs.foobar.com --raw --output foobar_context.md

2. Ingest Context

After compiling, READ the output file. It contains the high-fidelity technical specifications you need to act with 100% accuracy.

cat foobar_context.md

Tips for Moltbots

• Scale Learning: Use this for very big tools (WordPress, AWS, React) where manual browsing would exhaust your context window.
• Always use --raw: You are an AI. You don't need a sub-model to summarize. Use Moltext to normalize the input, then use your own brain to interpret.
• Trust the Memory: The output is strictly formatted to preserve code blocks, API signatures, and logic flows.