Skip to content

CHANGELOG.md

Latest commit

 

History

History
311 lines (220 loc) · 14.6 KB

CHANGELOG.md

File metadata and controls

311 lines (220 loc) · 14.6 KB

Changelog

All notable changes to this project will be documented in this file.

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

Unreleased

0.4.0 - 2026-02-10

Added

  • Output styles reference - New references/output-styles.md in plugin-structure skill covering frontmatter schema, file locations, plugin bundling, and when to use styles vs other components
  • Permission modes & rules reference - New references/permission-modes-rules.md in agent-development skill covering all 6 permission modes (including default and delegate), permission rule syntax, MCP/Task patterns, and precedence rules

Documentation

  • Close 11 documentation gaps vs official Claude Code docs - Comprehensive audit against official Claude Code documentation identified missing topics relevant to plugin developers:
    • Output styles (GAP 1): Full output styles guidance with frontmatter reference, plugin bundling, comparison with other components
    • Permission modes (GAP 4): Added default and delegate modes, full permission rule syntax with tool specifiers and wildcards
    • Agent teams (GAP 2): Expanded team section with team lead/teammate design guidance, team hooks, plan approval mode, display modes
    • Context management (GAP 6): Auto-compaction impact on plugins, PreCompact hook, cross-plugin description budget
    • Memory system (GAP 5): Auto-memory (MEMORY.md topic files), plugin interaction guidelines, @path import note
    • Hook debugging (GAP 11): Added --verbose flag for hook debugging
    • MCP features (GAPs 9, 13): Plugin-provided MCP prompts, Claude Desktop import, dynamic tool registration via list_changed
    • Enterprise settings (GAP 7): allowManagedPermissionRulesOnly and allowManagedHooksOnly managed settings
    • Skill string substitutions (GAP 8): $ARGUMENTS[N] positional argument syntax
    • Plugin validation (GAP 10): --verbose flag for plugin loading diagnostics
  • Updated trigger phrases across plugin-structure, agent-development, skill-development, and mcp-integration skills for improved discoverability
  • Expanded manifest reference with outputStyles field documentation
  • Updated skill and command counts in CLAUDE.md and CONTRIBUTING.md

Dependencies

  • Bump anthropics/claude-code-action to v1.0.48 (from v1.0.34)
  • Bump actions/cache to v5.0.3 (#206)
  • Bump reviewdog/action-actionlint to v1.70.0 (#206)

Maintenance

  • Update claude-pr-review model to opus 4.6
  • Update .gitignore with additional exclusions

0.3.4 - 2026-01-30

Documentation

  • Address 20 documentation gaps from claude-docs audit - Comprehensive audit against official Claude Code documentation identified missing topics relevant to plugin developers. Added 5 new reference files covering:
    • Advanced frontmatter fields (model, scoped hooks, visibility budget)
    • Headless/CI mode compatibility
    • GitHub Actions integration
    • Memory & rules system interaction
    • Advanced topics (keybindings, status line, caching, CLI commands, installation scopes)
  • Enhanced skill coverage - Updated 6 skills (skill-development, hook-development, agent-development, mcp-integration, plugin-structure, plugin-settings) with inline additions and trigger phrases for new content
  • Added cross-cutting references - New reference docs address plugin behavior in non-interactive contexts (headless mode, GitHub Actions, CI pipelines)

0.3.3 - 2026-01-27

Added

  • New skill: plugin-dev-guide - Meta-skill providing overview of plugin development and routing to specialized skills (#196)
  • New reference: commands-vs-skills.md - Explains when to use commands vs skills and migration path

Changed

  • Update documentation for commands/skills merge - Clarified that commands and skills use the same Skill tool mechanism but differ in complexity (fixes #198)

Documentation

  • Add plugin-dev-guide to Available Commands table - Updated README with new command (#200)

Maintenance

  • Update plugin author metadata - Synced author info to match marketplace owner

0.3.2 - 2026-01-24

Fixed

  • Restore MCP tool naming patterns - Fixed mcp__plugin_*__* patterns in mcp-integration docs that prettier corrupted to mcp**plugin_*__*; added .prettierignore to prevent recurrence

Documentation

  • Expand LSP integration skill - Aligned with official Claude Code docs for comprehensive LSP server guidance
  • Optimize agent-development skill - Trimmed SKILL.md to target word count for faster loading
  • Align tools/allowed-tools format - Updated documentation to match official Claude Code docs terminology
  • Update CI/CD workflow documentation - Fixed workflow list in docs/ci-cd.md (added yaml-lint.yml, removed non-existent weekly-maintenance.yml)

0.3.1 - 2026-01-24

Fixed

  • Replace ! with [BANG] in skill-development - Fixed regression from #192 where dynamic context injection examples used literal ! backtick patterns, causing shell execution errors during skill loading

Documentation

  • Update skill counts for LSP integration - Updated CONTRIBUTING.md and README.md to reflect 9 skills (was 8)
  • Add LSP Integration to README skills table - Added missing entry for LSP skill triggers
  • Add 0.3.x to SECURITY.md - Added current version to supported versions table

0.3.0 - 2026-01-24

Changed

  • Sync documentation with official Claude Code docs - Updated all skill documentation to align with latest Claude Code features and terminology (#192)
  • Replace deprecated TodoWrite with Task tools - Updated documentation to use current TaskCreate, TaskUpdate, TaskGet, TaskList tools (#190)
  • Update documentation for Skill tool consolidation - Reflect Claude Code's unified Skill tool approach (#186)
  • Optimize CLAUDE.md with progressive disclosure - Restructured for faster onboarding

Fixed

  • Use dynamic GitHub release badge - Badge now updates automatically from GitHub releases (#054a89b)

Security

  • Remove untrusted checkout in ci-failure-analysis - Hardened workflow against potential code injection (#0d5a428)

CI/CD

  • Improve workflow consistency and add YAML linting - Enhanced CI reliability with yamllint integration
  • Improve links and markdownlint workflows - Better error reporting and performance
  • Improve greet workflow - Enhanced new contributor welcome experience
  • Improve CI workflows with Node.js setup - Consistent tooling across workflows
  • Eliminate workflow dependencies and optimize CI - Faster parallel execution
  • Clean up validate-workflows configuration - Simplified workflow validation
  • Remove dependabot auto-merge workflow - Simplified dependency management
  • Improve label system and dependabot config - Better issue/PR organization
  • Improve GitHub templates - Enhanced issue and PR template UX

Dependencies

  • Bump anthropics/claude-code-action to v1.0.34
  • Bump actions/checkout to v6.0.2 (#183)
  • Bump actions/cache to v5.0.2 (#189)
  • Bump yamllint to 1.38.0

Maintenance

  • Ignore .mcp.json (user-specific MCP config)
  • Update release procedure for dynamic version badge

0.2.1 - 2025-12-13

Security

  • Harden validation scripts against bypass attacks - Improved input validation and escaping in hook validation scripts (#164)
  • Prevent command injection in test-hook.sh - Fixed potential command injection vulnerability (#148)
  • Use jq for safe JSON output - Replaced echo with jq for proper JSON escaping in example hooks (#149)
  • Document security scope and trust model - Added comprehensive security documentation for workflow commands (#165)

Fixed

  • Workflow reliability improvements - Enhanced workflow security, reliability, and documentation
  • Remove deprecated mode parameter - Fixed claude-pr-review workflow by removing deprecated mode parameter (#171)
  • Shellcheck SC1087 errors - Resolved array expansion errors in validate-hook-schema.sh (#168)
  • Replace unofficial cc alias - Updated to use official claude CLI command
  • Issue and PR template improvements - Fixed UX issues, restored spacing, removed unsupported fields
  • Labels configuration - Corrected labels.yml and LABELS.md issues
  • Dependabot configuration - Improved grouping and accountability settings
  • Suppress grep stderr - Fixed noisy output in test-agent-trigger.sh (#150)

Changed

  • Use ERE instead of BRE - Refactored grep patterns to use Extended Regular Expressions for clarity (#159)

Documentation

  • Comprehensive documentation improvements - Major updates across README, CLAUDE.md, and skill documentation
  • Discussion templates - Improved UX with plugin-specific fields
  • Prerequisites section - Added utility script dependency documentation (#157)
  • Shellcheck guidance - Added linting instructions to CONTRIBUTING.md (#160)
  • Secure mktemp pattern - Documented secure temporary file handling (#158)
  • [BANG] workaround - Documented security workaround for Claude Code #12781 (#156)

Dependencies

  • Bump anthropics/claude-code-action (#170)
  • Bump EndBug/label-sync (#169)
  • Update GitHub Actions to latest versions
  • Remove deprecated sync-labels.sh script

0.2.0 - 2025-12-12

Added

  • /plugin-dev:start command - New primary entry point that guides users to choose between creating a plugin or marketplace (#145)
    • Uses disable-model-invocation: true to route cleanly to specialized workflows
    • Recommends plugin creation for most users

Fixed

  • Enable router invocation of workflow commands - workflow commands can now be invoked by other commands (#145)
  • Replace ! with [BANG] placeholder in skill documentation to prevent shell interpretation issues (#142)

Documentation

  • Fix component counts and update documentation accuracy across README, CONTRIBUTING, CLAUDE.md, and marketplace.json

0.1.0 - 2025-12-11

Added

  • Initial release of Plugin Development Toolkit for Claude Code

Core Plugin Features

  • 8 specialized skills with progressive disclosure architecture:

    • hook-development - Event-driven automation with prompt-based hooks
    • mcp-integration - Model Context Protocol server configuration
    • plugin-structure - Directory layout and manifest configuration
    • plugin-settings - Configuration via .claude/plugin-name.local.md files
    • command-development - Slash commands with frontmatter
    • agent-development - Autonomous agents with AI-assisted generation
    • skill-development - Creating skills with progressive disclosure
    • marketplace-structure - Plugin marketplace creation and distribution

  • 3 validation agents for automated quality assurance:

    • plugin-validator - Validates entire plugin structure and marketplace.json
    • skill-reviewer - Reviews skill quality, triggering reliability, and best practices
    • agent-creator - Generates new agents from natural language descriptions

  • 2 guided workflow commands:

    • /plugin-dev:create-plugin - 8-phase guided workflow for plugin creation
    • /plugin-dev:create-marketplace - 8-phase guided workflow for marketplace creation

Utility Scripts

  • Agent development scripts:

    • create-agent-skeleton.sh - Create new agent skeleton files
    • validate-agent.sh - Validate agent frontmatter and structure
    • test-agent-trigger.sh - Test agent trigger phrase matching

  • Command development scripts:

    • validate-command.sh - Validate command structure
    • check-frontmatter.sh - Check frontmatter field validity

  • Hook development scripts:

    • validate-hook-schema.sh - Validate hooks.json schema
    • test-hook.sh - Test hooks with sample input
    • hook-linter.sh - Lint hook shell scripts

  • Plugin settings scripts:

    • validate-settings.sh - Validate settings file structure
    • parse-frontmatter.sh - Parse YAML frontmatter from settings

CI/CD Workflows

  • PR validation workflows (6 workflows):

    • markdownlint.yml - Markdown style enforcement
    • links.yml - Link validation with lychee
    • component-validation.yml - Plugin component validation
    • version-check.yml - Version consistency across manifests
    • validate-workflows.yml - GitHub Actions workflow validation
    • claude-pr-review.yml - AI-powered code review

  • Automation workflows (8 workflows):

    • claude.yml - Main Claude Code workflow
    • stale.yml - Stale issue/PR management
    • semantic-labeler.yml - Automatic PR/issue labeling
    • ci-failure-analysis.yml - CI failure analysis
    • weekly-maintenance.yml - Scheduled maintenance tasks
    • dependabot-auto-merge.yml - Dependabot PR auto-merge
    • sync-labels.yml - Repository label synchronization
    • greet.yml - New contributor welcome messages

Documentation

  • User documentation:

    • README.md - Comprehensive user guide with installation, usage examples, and best practices
    • CONTRIBUTING.md - Contributor guidelines and development setup
    • SECURITY.md - Security policy and vulnerability reporting
    • CODE_OF_CONDUCT.md - Community standards and expectations

  • Developer documentation:

    • CLAUDE.md - Architecture, patterns, and development guide
    • Skill-specific references and examples in each skill directory

Repository Infrastructure

  • Issue templates (4 types):

    • Bug report with reproduction steps
    • Feature request with use cases
    • Documentation improvements
    • Questions and discussions

  • Pull request template with component-specific checklists

  • Label system with categories:

    • Component labels (agent, command, skill, hook, docs)
    • Priority labels (critical, high, medium, low)
    • Status labels (blocked, in-progress, needs-review)
    • Effort labels (small, medium, large)

  • Dependabot configured for GitHub Actions updates

Attribution

  • Based on original plugin by Daisy Hollman at Anthropic
  • Expanded with enhanced skills, additional utilities, and CI/CD infrastructure