Context
Run artifacts now exist on disk, but users need a workspace to reopen them, retry failed handoffs, resume partial runs, and compare outputs.
Source: docs/WHISPERFORGE-MASTER-PLAN-2026-05-18.md (wf-run-workspace, Phase 8).
Linear: BC-229.
Acceptance Criteria
Tests/Evals
- Unit tests for manifest loading, partial-run status, and retry-safe operations.
- Rendered UI smoke if the run browser changes visible UI.
Verification
make test
venv/bin/python tests/ui_smoke.py
git diff --check
Agent Instructions
Build on existing .cache/runs manifests and history upserts. Treat destructive cleanup as out of scope.
Out of Scope
Do not implement arbitrary stage replay until reopen/retry is reliable.
Context
Run artifacts now exist on disk, but users need a workspace to reopen them, retry failed handoffs, resume partial runs, and compare outputs.
Source:
docs/WHISPERFORGE-MASTER-PLAN-2026-05-18.md(wf-run-workspace, Phase 8).Linear: BC-229.
Acceptance Criteria
Tests/Evals
Verification
make testvenv/bin/python tests/ui_smoke.pygit diff --checkAgent Instructions
Build on existing .cache/runs manifests and history upserts. Treat destructive cleanup as out of scope.
Out of Scope
Do not implement arbitrary stage replay until reopen/retry is reliable.