Migrate intrinsic quality criteria to .deepreview rules#249
Open
Migrate intrinsic quality criteria to .deepreview rules#249
Conversation
Enables agents to navigate back to prior steps when earlier outputs need revision. Clears all session progress from the target step onward, forcing re-execution to ensure consistency. Also teaches the new_job workflow about the iterative loop pattern so newly authored jobs can leverage go_to_step. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…step - Add JOBS-REQ-001.7 (go_to_step tool, 15 requirements) and renumber serialization section to JOBS-REQ-001.8 - Add JOBS-REQ-003.19 (step navigation state management, 10 requirements) - Update JOBS-REQ-003.7 to include go_to_step in session ID routing list - Update doc/architecture.md: add go_to_step to MCP tools list, diagram, tool docs, StateManager methods, and schemas listing - Add traceability comments to all 14 new test methods (9 in test_tools.py, 5 in test_state.py) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Catch StateError and raise ToolError with guidance on how to resume a workflow by restarting it and skipping steps with quality_review_override_reason. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Update JOBS-REQ-001.4.4: finished_step raises ToolError (not StateError) when no session exists, with guidance on resuming workflows - Add traceability for JOBS-REQ-001.7.14 to test_go_back_clears_subsequent_progress - Add test for JOBS-REQ-001.7.5: go_to_step with explicit session_id - Add test for JOBS-REQ-001.7.10: files on disk preserved after go_to_step - Add test for JOBS-REQ-001.7.11: concurrent entry navigates to first step Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Extract _build_active_step_info and _append_concurrent_info helpers in tools.py - Rename _resolve_session to resolve_session (public API) - Fix doc/mcp_interface.md: remove phantom field, add missing fields, remove obsolete config section - Fix doc/architecture.md: correct finished_step params, output format, StateManager/QualityGate signatures - Restructure SKILL.md dense paragraphs into numbered sub-lists - Extract Work Product Storage Guidelines to reference section in define.md - Standardize placeholder convention to square brackets in define.md - Move traceability comments before def lines in test_tools.py Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…w rules Intrinsic criteria (is the file well-formed?) should be enforced on ANY file change, not just during specific workflows. This extracts them into .deepreview rules and leaves only process criteria (did the agent follow the workflow?) in job.yml review blocks. New .deepreview rules: - job_yml_quality: step decomposition, review coverage, orphaned steps, promise tags - step_instruction_quality: completeness, specificity, examples, conciseness Enhanced deepreview_config_quality with valid-configuration and effective-instructions checks. Bumps deepwork_jobs 1.5.0→1.6.0 and deepwork_reviews 1.0.1→1.1.0. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…output formats - Fix 3 test methods silently skipped by pytest (testresolve_* → test_resolve_*) and convert from sync event loop to async (matching project convention) - Remove duplicate fixtures from 3 test classes (module-level ones suffice) - Update descriptions for deepreview_config_quality and step_instruction_quality to match their actual checks - Add PASS/FAIL output format sections to 4 rules that were missing them Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…n rename - Move 5 traceability comments in TestGoToStep from inside method body to immediately before the def line (matching project convention) - Update JOBS-REQ-003.7 spec: _resolve_session → resolve_session - Apply ruff format line-wrapping fix Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
.deepreviewrules so they run on ANY file change, not just during specific workflows.deepreviewrules:job_yml_quality(step decomposition, review coverage, orphaned steps, promise tags) andstep_instruction_quality(completeness, specificity, output examples, quality criteria, conciseness)deepreview_config_qualitywith valid-configuration and effective-instructions checksTest plan
uv run pytest— 708 tests passget_workflowsMCP call — all 4 jobs parse with 0 errorsget_configured_reviewsMCP call — 16 rules load including both new ones/reviewon the branch to validate the new rules fire correctly🤖 Generated with Claude Code