ST0019 - Add ash-expert: Modern Ash 3.0+ code quality specialist#4
Open
justtorrell wants to merge 103 commits intomainfrom
Open
ST0019 - Add ash-expert: Modern Ash 3.0+ code quality specialist#4justtorrell wants to merge 103 commits intomainfrom
justtorrell wants to merge 103 commits intomainfrom
Conversation
…e-v1.2.1 feat: Implement steel thread directory structure (v1.2.1)
- Remove obsolete scripts: fix_status.sh, organize_st.sh, update_frontmatter.sh - Integrate migrate_st_to_dirs functionality into stp_upgrade - Update STP_VERSION to 1.2.1 in main stp script - Update usage-rules.md prompt for directory structure - Regenerate usage-rules.md with v1.2.1 directory structure documentation - Standardize all scripts to use stp_* naming convention
The migrate_st_to_dirs script was consolidated into stp_upgrade. Updated tests to use the correct command with automatic confirmation for migration.
- Remove vestigial stp_st.bak backup file - Move RELEASE_NOTES_v1.2.1.md to docs/releases/1.2.1/RELEASE_NOTES.md - Create docs/releases directory structure for better organization - Update ST0010 info.md
…tories - Fixed logic to find all ST directories and organize based on status - Added status normalization (handles wip, WIP, in progress, etc.) - Prevents moving already organized directories - Added comprehensive tests for organize functionality The organize command now correctly: - Keeps WIP/In Progress threads in main directory - Moves Completed threads to COMPLETED/ - Moves Not Started threads to NOT-STARTED/ - Moves Cancelled threads to CANCELLED/
…d detection The stp init script was not creating the stp/.config/version file, which caused steel threads to not be detected properly in newly initialized projects. This was because STP v1.2.1 uses a directory-based structure for steel threads, but without the version file, the tools would fall back to looking for the old file-based format (ST*.md files). Changes: - Add STP_VERSION variable to stp_init script - Create stp/.config/version file during initialization with content 'stp_version: 1.2.1' - Version is now consistently set from the STP_VERSION variable This ensures that all new projects initialized with 'stp init' will have the correct version file and steel threads will be properly detected.
- Updated README.md with Intent branding throughout - Added comprehensive v2.0.0 entry to CHANGELOG.md - Updated DEPRECATIONS.md with STP → Intent deprecations - Created detailed release notes for v2.0.0 - All tests passing (86/86) This completes the Intent v2.0.0 release preparation.
- When Intent project is detected but lib/ doesn't exist, fall back to current directory - Fixes Ubuntu CI/CD where tests run in temp dirs within project hierarchy - Validates STARTDIR after all defaults are set, not during setting - Ensures fileindex works even when Intent project structure is incomplete
- Debug output will break tests but show us what's happening - Need to see why fileindex behaves differently on Ubuntu vs macOS - Will remove debug output after identifying the issue
- Debug directory validation and toggle mode checks - Track script execution flow to find where it exits with status 1 - Will help identify Ubuntu-specific issue
- Replace ((PROCESSED_COUNT++)) with PROCESSED_COUNT=1 - The ((0)) construct returns false/exit 1 when count is 0 - With set -e, this caused script to exit with status 1 - This is why tests failed only when no files matched the pattern - Also removed all debug output now that issue is fixed
- Add -C|--check flag to explicitly set files to checked [x] state - Add -U|--uncheck flag to explicitly set files to unchecked [ ] state - Keep existing -X|--toggle functionality for switching states - Add comprehensive tests for check/uncheck operations (14 new tests) - Update help documentation with new options and examples - Update CHANGELOG and release notes for v2.2.0 The check/uncheck operations are idempotent - checking an already checked file keeps it checked, making them more predictable than toggle for scripting and automation.
- Create new socrates agent implementing CTO Review Mode methodology - Facilitates Socratic dialog between CTO and Tech Lead personas - Helps with technical decision-making, architecture choices, and design reviews - Updates global agent manifest and documentation - Integrates with Intent project structure and steel threads The socrates agent transforms rubber-duck debugging into structured technical exploration, creating decision documentation that explains not just what was decided, but why.
- Replace hardcoded 'Installed: 2' checks with generic presence checks - Update tests to expect socrates agent in --all installation - Tests now check for 'Installation complete' and 'Installed:' instead of exact counts - This prevents tests from breaking when new agents are added
This release introduces centralized version management and significantly improves tool dependency checking to prevent silent failures. Major improvements: - Added VERSION file for single-source version management - Implemented get_intent_version() function for consistent version retrieval - Enhanced intent doctor with comprehensive tool dependency checking - Fixed silent failures when jq is missing during agent operations Tool dependency enhancements: - Categorized tools as required, core, and optional - Platform-specific installation instructions (macOS, Linux, WSL) - Clear error messages with actionable remediation steps - Better handling of tool alternatives (sha256sum vs shasum) Version management improvements: - All scripts now read from centralized VERSION file - Eliminated version inconsistencies across scripts - Simplified future version updates to single file change - Updated tests to use centralized version source Bug fixes: - Fixed silent failures in agent operations without jq - Resolved missing error messages for required dependencies - Corrected inadequate installation guidance for platforms - Fixed version number inconsistencies Technical details: - Enhanced jq error handling in intent_config and intent_agents - Improved fallback behavior when tools are missing - Better user experience with actionable error messages - Updated test framework to support centralized versioning This release ensures users are immediately aware of missing dependencies and provides clear guidance for resolution, eliminating frustrating silent failures and improving overall reliability.
- Add needs_v2_2_1_upgrade() function to detect v2.2.0 projects - Add migrate_v2_2_to_v2_2_1() function to handle the upgrade - Update upgrade logic to include 2.2.0→2.2.1 path - Fix upgrade sequencing to prevent version downgrades - Update release notes to document the fix This ensures all projects can upgrade to the latest version regardless of their current version.
- Updated `intent st start` to set status as "WIP" instead of "In Progress" - Maintains backward compatibility through existing status normalization - Updated tests to expect "WIP" status after starting steel threads - Aligns with project's preference for concise status terminology
- Modified intent_upgrade to check for existing CLAUDE.md before creating - Updated create_claude_md function to support optional overwrite parameter - Prevents loss of user customizations when running intent upgrade - Shows "Preserving existing CLAUDE.md..." when file exists - Shows "Creating CLAUDE.md..." only when file doesn't exist
- Comprehensive Worker-Bee specialist agent for Elixir applications - Interactive project structure mapping and validation - 6-layer architecture enforcement (Data, Functions, Tests, Boundaries, Lifecycles, Workers) - Mix tasks for scaffolding and compliance validation - Framework-agnostic support (Phoenix, OTP, libraries) - Custom agent creation documentation added
ST0018 - Add Worker-Bee Intent Agent for Architecture Enforcement
- Migrated worker-bee agent from legacy agents/ directory to new plugin architecture - Added worker-bee agent to intent/plugins/claude/subagents/ with resources directory - Updated version to 2.3.1 across all configuration files - Added complete upgrade paths from all versions to 2.3.1 - Added migrate_v2_3_0_to_v2_3_1() function for smooth upgrades - Updated CLAUDE.md with worker-bee agent documentation - Removed legacy agents/ directory structure The worker-bee agent provides Worker-Bee Driven Design (WDD) support for Elixir projects with 6-layer architecture enforcement, project structure mapping, and code scaffolding.
…ction - Added comprehensive antipattern detection to Elixir subagent - Detects and remediates 24 common Elixir antipatterns - Categories: Code (9), Design (6), Process (4), Meta-programming (5) - Full documentation in intent/plugins/claude/subagents/elixir/antipatterns.md - Antipatterns sourced from official Elixir documentation - Updated Elixir subagent (agent.md) with: - Antipattern detection and remediation workflows - Integration with systematic code review process - Example usage commands and report formats - Prevention principles and best practices - Version bump from 2.3.1 to 2.3.2 - Updated .intent/config.json and VERSION file - Added migrate_v2_3_1_to_v2_3_2() function - Updated upgrade paths in intent_upgrade script Run 'intent claude subagents sync' to update the Elixir subagent with the new capabilities.
The migrate_v2_3_1_to_v2_3_2 function was calling update_config_version which didn't exist. Added the helper function to properly update the version in .intent/config.json during migration.
Introduces a new specialized Claude subagent for comprehensive Ash Framework expertise, focusing on modern Ash 3.0+ patterns, code quality enforcement, and architectural guidance. Key capabilities: - Critical quality gates: Detect Ecto/Ash anti-patterns, validate resource definitions - Modern pattern promotion: Suggest bulk operations, atomic updates, domain-driven design - Performance optimization: Identify N+1 queries, recommend calculations and aggregations - Migration guidance: Help with ash_postgres patterns and constraint management - 4-tier expertise system from critical fixes to advanced transaction patterns The agent acts as a "strict but helpful mentor" to: - Enforce Ash principles and prevent common mistakes - Promote Ash 3.0+ features over legacy patterns - Provide concrete code examples and transformations - Reference Intent documentation at intent/docs/ref/ash/ Integrates with existing Intent agents (elixir, worker-bee) to provide specialized Ash framework knowledge while maintaining consistency with Worker-Bee Driven Design patterns and Elixir Usage Rules.
Complete documentation of the ash-expert agent creation process, including: - Design decisions and 4-tier expertise architecture - Implementation details and technical specifications - Comprehensive task breakdown (9 phases, 44 tasks) - Integration with Intent's agent ecosystem
eb63cac to
d44011e
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR introduces a new specialized Claude subagent for comprehensive Ash Framework
expertise, focusing on modern Ash 3.0+ patterns, code quality enforcement, and architectural
guidance - ST0019
Key Features
4-Tier Expertise System:
definitions, prevent runtime errors
domain-driven design
interfaces
Core Capabilities:
Repo.all,Repo.get, etc.)intent/docs/ref/ash/Integration
The ash-expert agent integrates seamlessly with existing Intent agents:
Files Changed
intent/plugins/claude/subagents/ash-expert/agent.md- Main agent definitionintent/plugins/claude/subagents/ash-expert/metadata.json- Agent metadataintent/plugins/claude/subagents/.manifest/global-agents.json- Registryintent/llm/AGENTS.md- Documentation with new agent descriptionintent/llm/AGENTS.md.bak- Backup of previous versionThis agent fills a critical gap in providing specialized Ash framework expertise while
maintaining consistency with Intent's development methodology and existing agent ecosystem.