Files
SuperClaude/.claude/commands/shared/auto-task-trigger.yml
NomenAK 7413eb631c Enhanced Task Management System - Adaptation of PR #5 (#7)
* Add comprehensive task management system with hybrid todo integration

- Automatic task detection based on complexity scoring (8+ points)
- Seamless .tasks/ folder structure (pending/in-progress/completed)
- Two-tier architecture: Tasks for persistence, TodoWrite for focus
- Auto-conversion of task breakdowns to TodoWrite steps
- Session recovery with context preservation
- Git integration with automatic branching
- Zero-interruption workflow for complex requirements

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

Co-Authored-By: Claude <noreply@anthropic.com>

* Enhance task management system with SuperClaude integration

- Update all .tasks/ paths to ./claudedocs/tasks/ for consistency
- Add plan mode integration for intelligent task suggestions
- Implement UltraCompressed format (~70% token reduction)
- Enhance git workflow with clean branch strategy and atomic commits
- Add task-ultracompressed.yml configuration
- Update install.sh to reflect new file counts
- Integrate with exit_plan_mode for complex work detection

This adapts PR #5's excellent task tracking functionality to fit
SuperClaude's architecture while adding valuable enhancements.

---------

Co-authored-by: Mohammad Aziz <mdaziz067@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
2025-06-24 11:29:02 +02:00

127 lines
3.6 KiB
YAML

# auto-task-trigger.yml - Seamless Automatic Task Mode
## Complexity Analysis Engine
```yaml
requirement_parser:
extract_keywords:
trigger_verbs: ["build", "create", "implement", "develop", "design", "add", "integrate"]
scope_nouns: ["system", "feature", "functionality", "component", "service", "application"]
complexity_flags: ["authentication", "database", "API", "full-stack", "architecture", "integration"]
scoring_system:
base_score: 0
trigger_verb: +2 points
scope_noun: +2 points
complexity_flag: +3 points
multi_step_indicator: +3 points
file_estimate: files * 1 point
thresholds:
auto_create: ≥8 points
suggest_create: 5-7 points
proceed_normal: <5 points
pattern_detection:
high_complexity_patterns:
- "build a * system"
- "create * authentication"
- "implement * database"
- "develop * application"
- "full-stack *"
- "end-to-end *"
multi_session_indicators:
- "complete *"
- "entire *"
- "comprehensive *"
- mentions of multiple technologies
- frontend + backend mentions
- database + API mentions
```
## Seamless Auto-Creation
```yaml
execution_flow:
1_silent_analysis:
- parse requirement in background
- calculate complexity score
- no user interruption
2_instant_decision:
score ≥8: auto_create_immediately
score 5-7: brief_notification_then_create
score <5: proceed_without_task
3_background_setup:
- generate task ID
- create task file
- setup git branch
- move to in-progress
- begin implementation
user_notification:
high_complexity: "[Task created: {id}] {brief_description}"
medium_complexity: "[Multi-step work detected - creating task] {brief_description}"
format: single_line_notification
timing: after_creation_before_work
no_interruption_policy:
- never ask "should I create a task?"
- never wait for confirmation
- never pause workflow
- seamless background operation
```
## Context Preservation
```yaml
auto_save_triggers:
context_threshold: >70% full
session_timeout: >30min inactive
error_recovery: on failures
manual_break: user requests pause
seamless_recovery:
startup_scan: check ./claudedocs/tasks/in-progress/
auto_resume: highest priority active task
context_restore: previous session state
notification: "Resuming: {task_title}"
session_continuity:
preserve_state: file paths, variables, decisions
track_progress: completed steps, current focus
handle_blockers: previous issues & solutions
maintain_context: architectural decisions
```
## Integration Patterns
```yaml
command_integration:
/user:build → auto-detect complexity → create task if needed → proceed
/user:implement → always create task → breakdown → execute
/user:create → analyze scope → task if multi-step → proceed
plan_mode_integration:
exit_plan_mode:
- analyze plan complexity
- count steps, files, technologies
- if complexity ≥8: "Create task to track this plan? (y/n)"
- if yes: create task with plan content
- track plan execution through task
planning_triggers:
- --plan flag detected
- risky operations (deploy, migrate)
- multi-phase work identified
- cross-cutting concerns
persona_activation:
architect: high complexity systems
frontend: UI/component requests
backend: API/database work
security: authentication/authorization
workflow_chains:
detect → create → breakdown → branch → implement → test → complete
requirement → task → steps → code → validation → merge
plan → analyze → suggest task → create → track → complete
```