mcpField
R
rust-learner

rust-learner

github.com/actionbookDevelopment1.6k110 installsUpdated 3w ago
Repository

About rust-learner

rust-learner is a skill for learning Rust by gathering current information about language features and crate updates. It prioritizes browser-fetcher agents and summarizes their results, with Fetch as a fallback if all agents fail.

Key features

  1. 01Fetches current Rust information
  2. 02Looks up crate pages through Actionbook
  3. 03Runs browser-fetcher agents in parallel
  4. 04Waits for all agents before fallback
  5. 05Summarizes retrieved results

Use cases

  1. 01Check a Rust crate's latest information
  2. 02Research Rust language features
  3. 03Summarize crate updates from web sources

MCP support

The workflow uses Actionbook MCP to find crate-page actions and retrieve URLs and selectors.

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/actionbook/actionbook --skill rust-learner

After installing, type/rust-learnerin Claude Code to confirm it loaded.

Details

From the project README

Rust Learner

Learn Rust by fetching real-time information about Rust language features and crate updates.

⚠️ Tool Priority & Waiting Rule

Priority:

  1. ✅ browser-fetcher agent (preferred)
  2. ⚠️ Fetch / WebFetch (only after all browser-fetcher agents fail)
  3. ⚠️ WebSearch (only when search engine results are needed)

⛔ No "Racing Ahead":

  • After launching browser-fetcher agents, MUST wait for ALL of them to complete
  • DO NOT use WebSearch/Fetch as "supplements" while waiting
  • Only use fallback tools after ALL browser-fetcher agents have failed

Workflow

Step 1: actionbook MCP
search_actions("lib.rs crate")  → get action ID
get_action_by_id(id)            → get URL and selectors
Step 2: Launch browser-fetcher agents
Launch multiple browser-fetcher agents in parallel
Step 3: Wait for ALL agents to complete
⛔ DO NOT use other tools during this time
✅ Wait for TaskOutput to return all results
Step 4: Summarize results
  • If agents succeed: summarize content for user
  • If ALL agents fail: use Fetch as fallback

Example

User: Query tokio latest version

✅ CORRECT:
1. Launch browser-fetcher: lib.rs/crates/tokio
2. Launch browser-fetcher: crates.io/crates/tokio
3. Wait for BOTH agents to complete
4. Summarize results

❌ WRONG:
1. Launch browser-fetcher agents
2. While waiting, use WebSearch("tokio latest")  ← Racing ahead!
3. Mix multiple result sources

Comments

Markdown supported