Changelog

All notable changes to rust-skills will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[2.0.9] - 2025-01-22

Changed

• Improved installation documentation in all READMEs (EN/ZH/JA):
  • Added two-step marketplace installation guide (/plugin marketplace add + /plugin install)
  • Clarified the difference between adding marketplace and installing plugin

Added

• Dependent Skills section in all READMEs:
  • actionbook/actionbook - MCP server for website action manuals
  • vercel-labs/agent-browser - Browser automation tool

---

[2.0.8] - 2025-01-22

Added

• 5 new LSP-based skills for code intelligence:
  • rust-code-navigator - Navigate code using LSP (goToDefinition, findReferences, hover)
  • rust-call-graph - Visualize function call hierarchies (incomingCalls, outgoingCalls)
  • rust-symbol-analyzer - Analyze project structure (documentSymbol, workspaceSymbol)
  • rust-trait-explorer - Explore trait implementations (goToImplementation)
  • rust-refactor-helper - Safe refactoring with impact analysis

Technical

• All LSP skills use allowed-tools: ["LSP", "Read", "Glob"] for safety
• Added comprehensive workflow documentation for each LSP operation

---

[2.0.7] - 2025-01-22

Added

• New skill: rust-deps-visualizer - Generate ASCII art dependency graphs
  • Support --depth and --features options
  • Size visualization and category grouping

Changed

• 22 reference skills now marked with user-invocable: false
  • m01-m07 (Layer 1: Language Mechanics)
  • m09-m15 (Layer 2: Design Choices)
  • domain-* (Layer 3: Domain Constraints)
  • coding-guidelines
• 5 command skills now have argument-hint for better UX
  • meta-cognition-parallel: <rust_question>
  • rust-skill-creator: <crate_name|std::module>
  • rust-daily: [today|week|month]
  • core-fix-skill-docs: [crate_name] [--check-only]
  • core-dynamic-skills: [--force] | <crate_name>
• 2 skills now have allowed-tools restrictions
  • unsafe-checker: Read, Grep, Glob only (no Bash for safety)
  • rust-learner: Task, Read, Glob only
• 2 skills now use dynamic context injection (!cmd``)
  • domain-embedded: Auto-injects .cargo/config.toml
  • m11-ecosystem: Auto-injects Cargo.toml dependencies

Documentation

• Skills now follow Claude Code Plugin Marketplace best practices

---

[2.0.6] - 2025-01-22

Fixed

• OpenCode installation - Rewrote installation guide with correct instructions config (Issue #6, thanks @DoiiarX)
  • Use instructions config option instead of non-existent plugins key
  • Restructured .opencode/ directory: plugin/ → instructions/
  • Added troubleshooting section and feature comparison table

---

[2.0.5] - 2025-01-22

Added

• Experimental: meta-cognition-parallel skill for three-layer parallel analysis
  • agents/layer1-analyzer.md - Language mechanics analysis agent
  • agents/layer2-analyzer.md - Design choices analysis agent
  • agents/layer3-analyzer.md - Domain constraints analysis agent
• New command: /meta-parallel for testing parallel meta-cognition

Documentation

• NPX installation note: recommend plugin installation for full functionality

---

[2.0.4] - 2025-01-22

Added

• context: fork support for task-based skills (Issue #4, thanks @pinghe)
  • rust-skill-creator - runs in isolated general-purpose agent
  • core-dynamic-skills - runs in isolated general-purpose agent
  • core-fix-skill-docs - runs in isolated general-purpose agent
  • rust-daily - runs in isolated Explore agent

Documentation

• Added analysis of which skills benefit from forked context

---

[2.0.3] - 2025-01-22

Fixed

• .mcp.json - Added missing mcpServers wrapper (PR #3)
• README.md, README-zh.md, README-ja.md - Added NPX installation method (PR #1)

Changed

• Updated feature comparison tables to include NPX method

---

[2.0.2] - 2025-01-22

Changed

• README.md, README-zh.md, README-ja.md - Added Marketplace installation method, version badge
• VERSION - Updated to 2.0.2
• metadata.json - Updated version, stats, skills list structure

---

[2.0.1] - 2025-01-22

Added

• tests/scenarios/domain-skills.md - Layer 3 domain skills test scenarios
• tests/scenarios/layer2-skills.md - Layer 2 design skills test scenarios

Changed

• hooks/hooks.json - Added domain keywords (fintech, kubernetes, embedded, IoT, ML, etc.)
• tests/README.md - Updated with complete test coverage summary

---

[2.0.0] - 2025-01-22

Added

Plugin Marketplace Support

• Added .claude-plugin/marketplace.json with official schema
• Simplified .claude-plugin/plugin.json (skills auto-discovered)
• Support for /plugin marketplace add actionbook/rust-skills

Domain Skills (Layer 3)

• domain-fintech: Financial technology patterns
• domain-web: Web development patterns
• domain-cli: CLI application patterns
• domain-embedded: Embedded systems patterns
• domain-cloud-native: Cloud-native patterns
• domain-iot: IoT patterns
• domain-ml: Machine learning patterns

Negotiation Protocol

• Added _meta/negotiation-protocol.md for comparative queries
• Added _meta/negotiation-templates.md for response formats
• Support for "compare", "vs", "best practice" queries

Changed

rust-router Optimization

• 56% context reduction (18.7 KB → 8.1 KB)
• Moved negotiation details to patterns/negotiation.md
• Moved workflow examples to examples/workflow.md
• Moved OS-Checker integration to integrations/os-checker.md
• Removed redundant skill file paths listing

Test Coverage

• Complete trigger test coverage (25/25 user-facing skills)
• Added Layer 2 skills tests (m05, m09-m15)
• Updated quick test commands

Fixed

• Fixed incorrect repository links (anthropics/rust-skills → actionbook/rust-skills)
• Fixed actionbook link (anthropics/actionbook → actionbook/actionbook)

Documentation

• Added docs/context-optimization.md with optimization details
• Updated docs/rust-skills-introduction.md

---

[1.0.0] - 2024-01-16

Added

Core Infrastructure

• Initial release of rust-skills Claude plugin
• 15 meta-question skills (m01-m15) for Rust learning
• rust-router for intelligent skill routing
• rust-learner for crate and version information

Skills

• m01-ownership: Memory ownership and lifetimes
• m02-resource: Resource management (Box, Rc, Arc)
• m03-mutability: Mutability patterns
• m04-zero-cost: Zero-cost abstractions
• m05-type-driven: Type-driven design
• m06-error-handling: Error handling patterns
• m07-concurrency: Concurrency and async
• m08-safety: Merged into unsafe-checker
• m09-domain: Domain modeling
• m10-performance: Performance optimization
• m11-ecosystem: Ecosystem integration
• m12-lifecycle: Resource lifecycle
• m13-domain-error: Domain error handling
• m14-mental-model: Mental model construction
• m15-anti-pattern: Anti-pattern recognition

unsafe-checker Skill

• 47 unsafe rules organized by category
• Checklists for writing and reviewing unsafe code
• FFI best practices and patterns
• Examples for safe abstractions

coding-guidelines Skill

• Compressed P rules by category (~80 core rules)
• G rules summary (g-compressed.md)
• Clippy lint mapping index

Agents

• crate-researcher: Fetch crate info from lib.rs/crates.io
• rust-changelog: Fetch Rust version changelog
• docs-researcher: Fetch API documentation
• clippy-researcher: Fetch Clippy lint information
• browser-fetcher: Generic web content fetching

Commands

• /guideline: Query coding guidelines
• /guideline --clippy: Map Clippy lints to rules
• /unsafe-check: Check file for unsafe issues
• /unsafe-review: Interactive unsafe code review

Deep Dive Content

• Common error patterns and fixes
• Lifetime patterns guide
• Async patterns guide
• Thread-based concurrency patterns
• Performance optimization guide
• Mental model construction guide
• Anti-pattern recognition guide
• Language comparison documents

Code Templates

• Error handling templates (thiserror, anyhow)
• Concurrency templates (worker-pool, async-task)
• FFI templates (safe-wrapper)

Testing

• Test scenarios for all major skills
• Validation script for skill configuration
• Agent integration test scenarios

Caching

• Agent result caching system
• Configurable TTL per category
• Cache management utilities

Technical Details

• Skills use YAML frontmatter for triggers
• Agents use haiku model for efficiency
• Routing based on keywords and error codes
• Chinese language triggers supported

[Unreleased]

Planned

• More code templates
• Index generation tools
• Additional dynamic skill support

你可能感兴趣的文章

Rust Skills - Claude Instructions

Rust Skills

Rust Skills Trigger Test Checklist

Rust Skills - Agent Instructions

Rust Skills Review Report

Rust Skills

Rust Skills