manual_slop/conductor/tracks/gencpp_dogfood_feedback_20260510/plan.md

Implementation Plan: GenCpp Dogfood Feedback Loop

Phase 1: Verify GenCpp Project Access

Focus: Ensure Manual Slop can properly target and load the gencpp project

• Task 1.1: Investigate how _allowed_paths is populated in mcp_client.py:configure() - determine if gencpp can be added as a secondary project
• Task 1.2: If access is possible, verify C:/projects/gencpp is accessible as a project root in Manual Slop
• Task 1.3: Check that .ai/manual_slop.toml or manual_slop.toml is recognized at the gencpp root
• Task 1.4: Confirm conductor directory resolves to .manual_slop/conductor relative to gencpp root
• Task 1.N: Write tests for project path resolution targeting external directories

Phase 2: Establish Feedback Capture Mechanism

Focus: Create a structured way to log issues found during dogfooding

• Task 2.1: Verify JOURNAL.md in the gencpp conductor directory is writable
• Task 2.2: Create a "Findings" section in the track state for capturing issues
• Task 2.3: Document the feedback workflow in track index.md
• Task 2.N: Write tests for findings capture mechanism

Phase 3: Validate Feedback Loop

Focus: Confirm the feedback mechanism works end-to-end

• Task 3.1: User reports an issue encountered while using Manual Slop with gencpp
• Task 3.2: Issue is captured in a structured format (findings log)
• Task 3.3: Issue can be converted into a new track for Manual Slop improvement
• Task 3.N: User manual verification of the complete feedback loop