tdd-task-progress-context-retainer

name	tdd-task-progress-context-retainer
description	Track resumable progress for TDD work by capturing red, green, and refactor state, exact test status, implementation progress, and next steps in a structured journal. Use when TDD spans interruptions, multiple files, multiple crates, or handoffs and Codex must preserve precise resume context.

Tdd Task Progress Context Retainer

Use this skill to preserve continuity in active TDD work and provide deterministic resume context across interruptions, handoffs, or multi-file sessions.

Workflow

1. Initialize a journal once per workstream.

• Run scripts/progress_journal_orchestrator.py init.
• Store one journal per feature or ticket.

2. Capture checkpoint updates continuously.

• Run scripts/progress_journal_orchestrator.py update after each meaningful change:
• Test batch created or changed.
• Phase transition (Red, Green, Refactor).
• Multi-file or multi-crate edits.
• Performance or error regression discovery.

3. Resume work from the last saved state.

• Run scripts/progress_journal_orchestrator.py snapshot.
• Use the latest checkpoint to restore current phase, failing tests, implementation status, and next action.

4. Emit handoff summaries.

• Produce a short human-readable status with:
• current phase
• top three next steps
• blockers
• unresolved design decisions

5. Resume from evidence, not memory.

• Before continuing after a pause, run snapshot and restate:
• exact failing tests
• files in motion
• why the next step is the next step
• any constraint that would make a naive continuation dangerous

Journal Contract

Capture these sections in every checkpoint:

• Current Phase
• Tests Written
• Implementation Progress
• Current Focus
• Next Steps
• Context Notes
• Performance/Metrics

Use explicit identifiers when available:

• test names
• file paths
• function signatures
• error messages

Script Usage

python3 scripts/progress_journal_orchestrator.py init \
  --path journals/feature-a.md \
  --task "Feature A end-to-end TDD"

python3 scripts/progress_journal_orchestrator.py update \
  --path journals/feature-a.md \
  --phase Red \
  --focus "Write failing tests for endpoint validation" \
  --test "test_endpoint_rejects_empty_payload|failing|expects 400 on empty payload" \
  --implementation "api/handlers.rs: added request parser scaffold" \
  --next-step "Implement minimum handler logic to satisfy validation tests" \
  --note "Need shared fixture helper for request setup"

python3 scripts/progress_journal_orchestrator.py snapshot \
  --path journals/feature-a.md

Guardrails

• Never leave Next Steps empty.
• Track why for major decisions, not only what.
• Record cross-component dependencies explicitly.
• Keep updates small and frequent to reduce context drift.
• Prefer append-only checkpoints over rewriting history.

Context Strategy

Load progressively:

1. Load this SKILL.md.
2. Load Journal schema and examples.
3. Load Checkpoint cadence playbook for long-running sessions.
4. Load Resume and handoff prompts when resuming or transferring work.

References

• Journal schema and examples
• Checkpoint cadence playbook
• Resume and handoff prompts