Tags: ClavixDev/Clavix
Tags
feat(v5.8.2): Clarifying Questions Protocol + Template Customization … …Guide ### Added - Clarifying Questions Protocol component with 95% confidence threshold - Integrated into AGENT_MANUAL for universal agent access - Applied to planning/improvement commands (improve, prd, plan, start, summarize) - Comprehensive template customization documentation ### Changed - Updated MANIFEST.md with clarifying-questions component - Version bumped to 5.8.2 across all templates
feat(v5.8.0): enhanced plan and verify templates with technical depth - plan.md: Added comprehensive context analysis phase and low-level engineering task generation - verify.md: Transformed into spec-driven technical audit with senior code review approach - Improved error handling protocols and self-correction mechanisms - Enhanced implementation specificity with exact file paths and technical constraints 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <[email protected]>
chore: release v5.7.1 - improved template consistency and tone - Bump version to 5.7.1 - Update Agent Transparency version across all canonical templates - Add comprehensive template style guide (docs/template-style-guide.md) - Add validation script for template consistency (scripts/validate-consistency.ts) - Fix validation script to use tsx instead of ts-node for Node v25 compatibility Major improvements to verify.md: - Rewrite with friendly, conversational tone matching other commands - Simplify complex tables and verification matrices - Add clearer examples and better structured reporting - Update intro: 'Check Your Work' instead of formal technical language - Improve troubleshooting with 'Don't Panic' approach Major improvements to refine.md: - Rewrite with more approachable, casual tone - Simplify workflow sections and remove formal checkpoints - Add better conversational examples - Streamline PRD and prompt refinement flows - Improve 'When Things Go Wrong' troubleshooting section Minor improvements to improve.md: - Add new 'What This Does' section explaining the workflow - Detail smart depth selection logic (quality score-based) - Clarify 6-dimension assessment approach All changes maintain functionality while significantly improving user experience and consistency across the command suite.
feat(v5.7.0): template quality, schema cleanup, consistency tests Template Improvements: - Standardized section order across all 9 templates (CLAVIX MODE → Self-Correction → State Assertion) - verify.md: Present report in chat, ask user what to do next - plan.md: Added verification section after plan generation - archive.md: Explicit tool clarity (Read, Bash/Move, etc.) - Added state markers for workflow continuity to improve, prd, plan, summarize templates - Added recovery patterns to self-correction-protocol.md Schema Cleanup: - Removed unused supportsFrontmatter from all adapter types and tests - Documented argument placeholder strategy in docs/templates.md Documentation: - Added OUT OF SCOPE section to CONTRIBUTING.md (MCP, Git hooks, CI/CD) - Fixed CHANGELOG reference in PR checklist - Fixed LLXPRT naming (was LLXpert) in integrations.json and docs Testing: - Added naming-consistency.test.ts (8 tests) - All 1311 tests passing 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <[email protected]>
feat(v5.6.6): template quality improvements & test coverage ### Added - 21 new CLI tests (10 update.ts, 11 init.ts reconfiguration flows) - Task dependency guidance in plan.md with explicit markers - Scope creep detection and handling in start.md - Archive size management guidance in archive.md - Multi-topic handling in refine.md troubleshooting - Workflow-specific dimension usage in quality-dimensions.md ### Fixed - Package.json description now lists all 9 slash commands - Fixed "19+" to "20" AI tools count in docs - Standardized quality dimension terminology in refine.md - Added /clavix:refine to MANIFEST.md usage matrix Total: 1306 tests passing 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <[email protected]>
feat(v5.6.5): test expansion, init refactoring, JSON schemas Improvements: - Add 30 logger utility tests (tests/utils/logger.test.ts) - Expand update.test.ts from 39 to 84 tests (+45 tests) - Add conflicting flags validation to update command - Add /clavix:refine to octo.md, warp.md agent templates - Add clean script to package.json for build cleanup Refactoring: - Extract init.ts into 6 helper modules (src/cli/helpers/init/) - directories.ts, config.ts, documentation.ts - commands.ts, integrations.ts, prompts.ts New features: - Generate JSON schemas from Zod (schemas/ directory) - config.schema.json for .clavix/config.json validation - integrations.schema.json for integrations.json validation - Add scripts/generate-json-schema.ts Test count: 1269 tests passing 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <[email protected]>
fix(v5.6.4): UserConfigSchema missing fields Fixed Zod schema to match ClavixConfig interface: - Added version, templates, outputs, preferences, experimental fields - Eliminates false "Unknown fields" warnings during clavix update 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <[email protected]>
feat(v5.6.3): Zod validation, universal adapters, template standardiz…
…ation
Added:
- Zod schema validation for integrations.json and config.json
- 4 universal adapters (agents-md, copilot-instructions, octo-md, warp-md)
- `type` field in integrations.json ("standard" or "universal")
- docs/templates.md authoring guide
- 23 schema validation tests
Changed:
- Summarize→Plan workflow: project name prompt, fallback detection
- Error output standardized to this.log()/this.error() in CLI
- All 9 template headers use "## State Assertion (REQUIRED)"
- README.md reorganized by workflow with Mermaid diagram
- CONTRIBUTING.md with quick-start decision table
Fixed:
- JSON parsing crashes with try-catch protection
- Test suite for adapter count changes (16 → 20)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <[email protected]>
PreviousNext