Refactor job.yml format: inline steps, step_arguments, DeepWork Reviews#256
Draft
nhorton wants to merge 16 commits intoclaude/remove-runner-arg-oc2v7from
Draft
Refactor job.yml format: inline steps, step_arguments, DeepWork Reviews#256nhorton wants to merge 16 commits intoclaude/remove-runner-arg-oc2v7from
nhorton wants to merge 16 commits intoclaude/remove-runner-arg-oc2v7from
Conversation
…ws quality gate
Unify the job.yml format around workflows as the primary structure with inline
steps, replacing the old two-level format (root steps[] + workflows referencing
step IDs) and the bespoke ClaudeCLI quality gate.
Key changes:
- step_arguments[] define shared data vocabulary (string/file_path) flowing between steps
- workflows{} are objects with inline steps (no separate step .md files)
- Quality gate uses DeepWork Reviews infrastructure instead of Claude CLI subprocess
- review blocks on step_arguments/outputs create dynamic ReviewRules at runtime
- process_quality_attributes review work_summary against criteria
- json_schema on step_arguments validates file outputs before reviews
- notes → work_summary, current_entry_index → current_step_index
- Removed: version field, root-level steps[], dependencies, hooks, exposed/hidden,
instructions_file, concurrent step groups, ClaudeCLI module, --external-runner flag
All 3 standard jobs and library job migrated with instructions inlined.
627 tests pass, ruff clean, mypy clean.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Fix status.py to use dict-based workflows API and WorkflowStep.name - Fix entry_index -> step_index in test_state.py and test_status.py - Fix ruff import ordering and formatting across rebased files Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
nhorton
force-pushed
the
job_yml_refactor
branch
from
fe4fa3a to
a47898d
Compare
The release process pins the version in .mcp.json args (e.g. "deepwork==0.10.0a1"), which broke the exact-match assertion. Use startswith instead so the test passes for both unpinned and version-pinned args. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Rewrite CONTRIBUTING.md pre-release section to lead with uv tool install (what the MCP server actually uses via uvx), not uv pip install - Add --platform claude to all serve command examples in CONTRIBUTING.md and doc/architecture.md to match the actual .mcp.json config - Remove lowercase claude.md symlink (duplicate of CLAUDE.md, causes collisions on case-insensitive filesystems) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This reverts commit bd70adf.
…nt fixes - Sync __init__.py version with pyproject.toml (0.7.0 → 0.9.8) - Add __init__.py version update to prepare-release workflow - Fix doc/mcp_interface.md: notes→work_summary, add inputs param, fix ActiveStepInfo/ExpectedOutput types, add post_workflow_instructions - Fix doc/architecture.md: needs_review→needs_work, .mcp.json args - Add engineer job to library/jobs/README.md - Fix mypy type annotations in test_state.py and test_tools.py Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* Fix MCP version test to allow release-pinned package names The release process pins the version in .mcp.json args (e.g. "deepwork==0.10.0a1"), which broke the exact-match assertion. Use startswith instead so the test passes for both unpinned and version-pinned args. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Fix: update marketplace.json version for pre-releases too The marketplace version update was gated to stable-only, so pre-release builds pushed to the pre-release branch had a stale version in .claude-plugin/marketplace.json. This meant users installing via /plugin marketplace add ...#pre-release got mismatched versions. Changes: - Remove the is_prerelease == false guard on the marketplace update step - Use plugin_version (semver) instead of pypi_version for consistency with plugin.json (no behavior change for stable where they're equal) - Add marketplace.json to the pre-release git add Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Update test for mcp_command requirement validation * Clarify requirement in test_mcp_command_is_uvx_deepwork_serve Updated comment to clarify requirement for test. * Fix ruff formatting in test_claude_plugin.py Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add 36 traceability comments (THIS TEST VALIDATES A HARD REQUIREMENT) back to test_tools.py and test_quality_gate.py, mapping tests to their JOBS-REQ-001.* requirement IDs. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Update JOBS-REQ-001/002/004 specs to match new job format and Reviews-based quality gate - Deprecate JOBS-REQ-009 (claude_cli.py was deleted) - Remove legacy instance_id from skill files and post_compact hook - Update .github/workflows/README.md with __init__.py and marketplace.json version bumps, remove duplicate sections Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When MCP tool interfaces change (tools.py, schemas.py, server.py), automatically review skill files, hooks, and platform content to catch stale parameter names or outdated behavior descriptions. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Rewrite stale sections in doc/mcp_interface.md: output types, response fields, quality gate behavior, review types, CLI options, go_to_step - Fix JOBS-REQ-001.1.13 → JOBS-REQ-001.1.8 in test_server.py - Add 19 traceability comments to test_quality_gate.py for JOBS-REQ-004 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…ents - JOBS-REQ-001.5.8 → JOBS-REQ-001.5.4 (file_path list validation) - JOBS-REQ-001.6.8 → JOBS-REQ-001.6.5, JOBS-REQ-001.6.6 (abort nested) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…ith RFC 2119 semantics Review instructions now use RFC 2119 keywords (MUST, SHOULD, MAY) to determine pass/fail severity. MUST/SHALL violations cause failure; SHOULD/RECOMMENDED violations fail only when easily achievable; other requirements produce feedback without failure. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…cleanup - Update job.yml.template and job.yml.example to new format (step_arguments, workflows object, inline instructions, process_requirements with RFC 2119) - Restore shared_jobs workflow with inlined sync_shared_jobs instructions - Fix deprecated `deepwork install` reference to mention standard jobs - Move inline import to top-level in quality_gate.py - Remove unused `step` param from _collect_output_file_paths - Expand .deepreview rules to cover doc/job_yml_guidance.md and library jobs Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…irements Convert engineer, platform_engineer, repo, and research jobs from old format (version, root-level steps[], instructions_file, dependencies, old reviews) to new format (step_arguments, workflows object, inline instructions, process_requirements with RFC 2119 keywords, output-level review blocks). Also fix stale CLAUDE.md section about standard job file structure. Co-Authored-By: Claude Opus 4.6 (1M context) <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
steps[]+ workflows referencing step IDs)step_arguments[]as shared data vocabulary (string/file_path) flowing between steps with optional review blocks and json_schema validationsteps/*.mdfiles)Test plan
🤖 Generated with Claude Code