Epic Overview
Implement a workflow enforcement system that intercepts direct code edits and proactively suggests @simple-mode workflows, ensuring LLMs act as "Senior Developers" by default.
Goal
- Workflow usage rate: 95%+ (currently ~30%)
- Issue detection: 95%+ caught before user sees code
- Performance: <50ms interception latency
- Test coverage: ≥85%
Stories (Implementation Order)
Total: 8 story points | Timeline: 2 weeks
Reference
- Story File:
stories/enh-001-workflow-enforcement.md
- Architecture:
docs/architecture/enh-001-workflow-enforcement-architecture.md
- API Spec:
docs/api/enh-001-api-spec.md
- Enhanced Prompt:
.tapps-agents/sessions/enh-001-enhanced-prompt.md
Workflow Commands
Session 1 (Issue #10):
@simple-mode *build "ENH-001-S4: Configuration System - Load workflow enforcement configuration from .tapps-agents/config.yaml with validation. Create EnforcementConfig dataclass with mode (blocking/warning/silent), confidence_threshold (default 60), suggest_workflows, block_direct_edits. Implement from_config_file() class method with YAML parsing, schema validation, and defaults. File: tapps_agents/core/llm_behavior.py (150 lines). Closes #10"
Session 2 (Issue #11):
@simple-mode *build "ENH-001-S2: Intent Detection System - Classify user intent into workflow types (*build, *fix, *refactor, *review) with confidence scoring. Keyword-based detection with 4 keyword sets. Calculate confidence 0-100%, trigger at ≥60%. File: tapps_agents/workflow/intent_detector.py (150 lines). Performance: <5ms latency. Closes #11"
Session 3 (Issue #12):
@simple-mode *build "ENH-001-S3: User Messaging System - Format enforcement messages for blocking and warning modes. Include detected intent, recommended workflow command, benefits list, and override instructions. Support CLI and IDE output formats. File: tapps_agents/workflow/message_formatter.py (100 lines). Closes #12"
Session 4 (Issue #13):
@simple-mode *build "ENH-001-S1: Core Workflow Enforcer - Coordinate enforcement workflow: intercept file operations, detect intent, apply policy, format message. Integrate with EnforcementConfig, IntentDetector, MessageFormatter. Hook into Write/Edit tools. File: tapps_agents/workflow/enforcer.py (250 lines). Performance: <50ms latency. Closes #13"
Epic Overview
Implement a workflow enforcement system that intercepts direct code edits and proactively suggests @simple-mode workflows, ensuring LLMs act as "Senior Developers" by default.
Goal
Stories (Implementation Order)
Total: 8 story points | Timeline: 2 weeks
Reference
stories/enh-001-workflow-enforcement.mddocs/architecture/enh-001-workflow-enforcement-architecture.mddocs/api/enh-001-api-spec.md.tapps-agents/sessions/enh-001-enhanced-prompt.mdWorkflow Commands
Session 1 (Issue #10):
Session 2 (Issue #11):
Session 3 (Issue #12):
Session 4 (Issue #13):