Commit Graph

5 Commits

Author SHA1 Message Date
rob ca73528068 feat: extend ai config with codex and gemini fallbacks 2025-11-01 14:59:45 -03:00
rob 4680f4b629 build: bundle shared ai config into installer 2025-11-01 14:44:25 -03:00
rob 4176f51e7d fix: Add YAML syntax fix and mock AI script for testing
- Fix missing space after colon in features.ai-rules.yml
- Add tools/mock_ai.sh for testing automation without real AI
- Ensures installer has valid YAML templates

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-10-31 09:18:59 -03:00
rob 5944269ed8 refactor: Support CLI-based AI providers (Claude Code, Gemini, Codex)
Major refactoring to support flexible AI provider configuration instead of
requiring direct API access. Users can now use whatever AI CLI tool they have
installed (claude, gemini, codex, etc.) without API keys.

## Changes to automation/agents.py

**New Functions:**
- `get_ai_config()` - Reads config from env vars or git config
  - Environment: CDEV_AI_PROVIDER, CDEV_AI_COMMAND (highest priority)
  - Git config: cascadingdev.aiprovider, cascadingdev.aicommand
  - Default: claude-cli with "claude -p '{prompt}'"

- `call_ai_cli()` - Execute AI via CLI command
  - Passes content via temp file to avoid shell escaping
  - Supports {prompt} placeholder in command template
  - 60s timeout with error handling
  - Parses JSON from response (with/without code blocks)

- `call_ai_api()` - Direct API access (renamed from call_claude)
  - Unchanged functionality
  - Now used as fallback option

- `call_ai()` - Unified AI caller
  - Try CLI first (if configured)
  - Fall back to API (if ANTHROPIC_API_KEY set)
  - Graceful failure with warnings

**Updated Functions:**
- `normalize_discussion()` - calls call_ai() instead of call_claude()
- `track_questions()` - calls call_ai() instead of call_claude()
- `track_action_items()` - calls call_ai() instead of call_claude()
- `track_decisions()` - calls call_ai() instead of call_claude()

**Configuration Precedence:**
1. Environment variables (session-scoped)
2. Git config (repo-scoped)
3. Defaults (claude-cli)

## Changes to docs/AUTOMATION.md

**Updated Sections:**
- "Requirements" - Now lists CLI as Option 1 (recommended), API as Option 2
- "Configuration" - Complete rewrite with 5 provider examples:
  1. Claude CLI (default)
  2. Gemini CLI
  3. OpenAI Codex CLI
  4. Direct API (Anthropic)
  5. Custom AI command
- "Troubleshooting" - Added "AI command failed" section, updated error messages

**New Configuration Examples:**
```bash
# Claude Code (default)
git config cascadingdev.aicommand "claude -p '{prompt}'"

# Gemini
git config cascadingdev.aiprovider "gemini-cli"
git config cascadingdev.aicommand "gemini '{prompt}'"

# Custom
git config cascadingdev.aicommand "my-ai-tool --prompt '{prompt}' --format json"
```

## Benefits

1. **No API Key Required**: Use existing CLI tools (claude, gemini, etc.)
2. **Flexible Configuration**: Git config (persistent) or env vars (session)
3. **Provider Agnostic**: Works with any CLI that returns JSON
4. **Backward Compatible**: Still supports direct API if ANTHROPIC_API_KEY set
5. **User-Friendly**: Defaults to "claude -p" if available

## Testing

-  get_ai_config() tests:
  - Default: claude-cli with "claude -p '{prompt}'"
  - Git config override: gemini-cli with "gemini '{prompt}'"
  - Env var override: codex-cli with "codex '{prompt}'"
-  extract_mentions() still works (no AI required)
-  All 6 workflow tests pass

## Impact

Users with Claude Code installed can now use the automation without any
configuration - it just works! Same for users with gemini or codex CLIs.
Only requires git config setup if using non-default command.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-10-30 18:31:27 -03:00
rob 72577bd30d feat: Implement Phase 2 - AI-powered discussion automation
Complete implementation of the AI-enhanced automation system with 5 specialized agents:

## New Files

1. **automation/agents.py** (270 lines)
   - Agent 1: normalize_discussion() - Extracts structured info from natural language
   - Agent 2: track_questions() - Identifies questions and tracks answers (OPEN/PARTIAL/ANSWERED)
   - Agent 3: track_action_items() - Manages TODO → ASSIGNED → DONE lifecycle
   - Agent 4: track_decisions() - Logs ADR-style decisions with rationale
   - extract_mentions() - Parses @Name and @all mentions (no API required)
   - call_claude() - Anthropic API integration with JSON response parsing

2. **automation/summary.py** (289 lines)
   - update_marker_block() - Non-destructive marker block updates
   - format_votes_section() - Formats vote counts and participants
   - format_questions_section() - Formats open/partial/answered questions
   - format_action_items_section() - Formats TODO/ASSIGNED/DONE tasks
   - format_decisions_section() - Formats ADR-style decisions
   - format_awaiting_section() - Groups @mentions by target
   - update_summary_file() - Orchestrates all section updates

3. **docs/AUTOMATION.md** (361 lines)
   - Complete system documentation
   - Phase 1 (basic) and Phase 2 (AI) explanations
   - Conversation guidelines and markers
   - Configuration and troubleshooting

## Enhanced Files

**automation/workflow.py** (enhanced to 349 lines, +132 lines)
- get_discussion_changes() - Git diff extraction for incremental processing
- update_summary_votes() - Marker block update for votes (Phase 1 fallback)
- process_discussion_with_ai() - Orchestrates all AI agents
- Enhanced _run_status() - Processes with AI when available, graceful fallback
- Dual import style for different execution contexts (pre-commit hook vs direct)

## Features

### Phase 1 (Always Enabled):
-  Vote parsing and tracking (READY/CHANGES/REJECT)
-  Latest vote per participant logic
-  Summary file auto-update and staging
-  Non-blocking git hook integration

### Phase 2 (Optional, requires ANTHROPIC_API_KEY):
-  @Mention extraction and tracking
-  Question identification and status tracking
-  Action item lifecycle management
-  Decision logging with rationale (ADR-style)
-  Incremental processing via git diff
-  Graceful degradation when API unavailable

## Architecture

- Modular design with separate concerns (agents, summary, workflow)
- Non-blocking operation (always exits 0)
- Dual-phase capability (basic + AI-enhanced)
- Graceful fallbacks at every level
- Comprehensive error handling

## Testing

- All 11 existing tests passing
- End-to-end testing in /tmp/cdev-test-phase2a confirmed:
  - Vote tracking works correctly
  - @Mentions extracted and grouped by target
  - Summary files updated non-destructively
  - Auto-staging of updated summaries
  - Pre-commit hook integration functional

## Documentation

- AUTOMATION.md covers architecture, usage, and troubleshooting
- Code comments explain complex logic
- Example formats provided for all features
- Configuration and setup instructions included

This completes the Phase 2 automation milestone, providing a sophisticated
AI-powered discussion summarization system while maintaining full backward
compatibility with basic vote-only mode.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-10-30 18:12:48 -03:00