feat(execution): wire Execute→Review→Simplify→Review pipeline into executor workflow by maystudios · Pull Request #20 · maystudios/maxsimcli

maystudios · 2026-03-02T04:52:05Z

Summary

Add post-task simplification pass (duplication, dead code, complexity checks) to execute-plan workflow and executor agent
Add post-wave 2-stage code review gate (spec compliance + code quality) with retry limits
Update executor agent flow to Execute → Simplify → Verify → Commit cycle with progress tracking table (EXEC-02) and task-based context loading (EXEC-03)
Add skill_paths to ExecutePhaseContext in init.ts so orchestrators know where to find skills
Update skill path references from .agents/skills/ to ~/.claude/skills/ across executor and orchestrator
Add wave review aggregation and progress tables to execute-phase orchestrator

Test plan

Unit tests pass (npm test — 54 tests, 2 suites)
Build succeeds (npm run build)
Lint passes (biome check)
Pack test confirms dist contents

🤖 Generated with Claude Code

…ecutor workflow - Add post-task simplification pass to execute-plan workflow (duplication, dead code, complexity) - Add post-wave 2-stage code review gate (spec compliance + code quality) - Update executor agent with Execute→Simplify→Verify→Commit cycle - Add orchestrator status tracking table (EXEC-02) with stage progression - Add task-based context loading instructions (EXEC-03) - Update skill paths from .agents/skills/ to ~/.claude/skills/ - Add skill_paths to ExecutePhaseContext in init.ts - Add progress table and wave review aggregation to execute-phase orchestrator Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Copilot

Pull request overview

This PR updates the execute-phase / executor workflow templates to add a post-task simplification step and a post-wave two-stage review gate, while also migrating skill directory references to ~/.claude/skills/ and exposing skill path info via the CLI init context.

Changes:

Add a per-task “simplify” pass (duplication/dead code/complexity) before committing, with re-verification.
Add a post-wave, two-stage review gate (spec compliance + code quality) and progress/review tables in execution reporting.
Update skill directory references and add skill_paths to ExecutePhaseContext output from cmdInitExecutePhase.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 7 comments.

File	Description
templates/workflows/execute-plan.md	Adds simplify pass and a post-wave code review gate to the execute-plan workflow template.
templates/workflows/execute-phase.md	Updates orchestrator wave flow to include a wave review step, progress tables, and skill path reference changes.
templates/agents/maxsim-executor.md	Updates executor guidance for task-scoped context loading, Execute→Simplify→Verify→Commit cycle, progress tracking, and skill path references.
packages/cli/src/core/init.ts	Extends the execute-phase init context with a skill path field.

Comments suppressed due to low confidence (1)

templates/agents/maxsim-executor.md:33

This section calls them “Project skills” but instructs checking ~/.claude/skills/ first and only then .claude/skills/ as a fallback. That ordering conflicts with later text about project skills overriding built-in/global skills. Clarify precedence (typically project-local overrides global) and align the check order + wording accordingly.

**Project skills:** Check `~/.claude/skills/` directory if it exists (also check `.claude/skills/` in the project root as a fallback):
1. List available skills (subdirectories)
2. Read `SKILL.md` for each skill (lightweight index ~130 lines)
3. Load specific `rules/*.md` files as needed during implementation
4. Do NOT load full `AGENTS.md` files (100KB+ context cost)
5. Follow skill rules relevant to your current task

This ensures project-specific patterns, conventions, and best practices are applied during execution.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-03-02T04:56:43Z

templates/workflows/execute-phase.md

+   # Get all files changed in this wave
+   WAVE_FIRST_COMMIT=$(git log --oneline --all --grep="{phase}-{first_plan_in_wave}" --reverse | head -1 | cut -d' ' -f1)


The suggested command uses --grep="{phase}-{first_plan_in_wave}", but first_plan_in_wave isn’t defined anywhere else in this workflow template (unlike {phase}-{plan} above). As written, this is likely to be copy-pasted and fail or produce an empty commit hash. Consider computing the first plan ID for the wave explicitly (or using the collected plan IDs/commit hashes from the wave) and reference that variable here.

Suggested change

# Get all files changed in this wave

WAVE_FIRST_COMMIT=$(git log --oneline --all --grep="{phase}-{first_plan_in_wave}" --reverse | head -1 | cut -d' ' -f1)

# Get all files changed in this wave.

# Assumes $WAVE_FIRST_COMMIT was recorded when starting the wave (the first commit for this wave).

Copilot · 2026-03-02T04:56:43Z

templates/workflows/execute-phase.md

       - .planning/config.json (Config, if exists)
       - ./CLAUDE.md (Project instructions, if exists — follow project-specific guidelines and coding conventions)
-       - .agents/skills/ (Project skills, if exists — list skills, read SKILL.md for each, follow relevant rules during implementation)
+       - ~/.claude/skills/ (Skills, if exists — list skills, read SKILL.md for each, follow relevant rules during implementation)


<files_to_read> now lists only ~/.claude/skills/, but the executor template also mentions a project-local .claude/skills/ fallback/override. If project-local skills are supported, the orchestrator’s startup context list should include them as well so executors discover the correct skill set consistently.

Suggested change

- ~/.claude/skills/ (Skills, if exists — list skills, read SKILL.md for each, follow relevant rules during implementation)

- ./.claude/skills/ (Project-local skills, if exists — override or extend global skills; list skills and read SKILL.md for each)

- ~/.claude/skills/ (Global skills, if exists — list skills, read SKILL.md for each, follow relevant rules during implementation)

Copilot · 2026-03-02T04:56:43Z

templates/agents/maxsim-executor.md

+1. Call `skill-context` or read the plan to get the task's file list
+2. Use the `Read` tool to load only those specific files
+3. If the task has no `Files:` field, load files referenced in the task description


task_context_loading instructs “Call skill-context” to get a task’s file list, but there’s no other reference in the repo to a skill-context command/tool. If this is meant to be a MAXSIM tool, it should be named consistently with other documented commands (e.g., maxsim-tools.cjs ...) or replaced with explicit instructions to parse the Files: field from the plan.

Suggested change

1. Call `skill-context` or read the plan to get the task's file list

2. Use the `Read` tool to load only those specific files

3. If the task has no `Files:` field, load files referenced in the task description

1. Read the plan (e.g., `PLAN.md`) and parse the current task's `Files:` field to get the task's file list

2. Use the `Read` tool to load only those specific files

3. If the task has no `Files:` field, load files explicitly referenced in the task description

Copilot · 2026-03-02T04:56:43Z

packages/cli/src/core/init.ts

  state_path: string;
  roadmap_path: string;
  config_path: string;
+  skill_paths: string;


ExecutePhaseContext is exported from the published CLI package; adding a new required field (skill_paths) is a breaking TypeScript change for any external consumers that construct this type. Consider making it optional (or adding a new optional field) to preserve compatibility, especially since it’s not yet used elsewhere in the codebase.

Suggested change

skill_paths: string;

skill_paths?: string;

Copilot · 2026-03-02T04:56:44Z

packages/cli/src/core/init.ts

@@ -431,6 +434,7 @@ export function cmdInitExecutePhase(cwd: string, phase: string | undefined, raw:
    state_path: '.planning/STATE.md',
    roadmap_path: '.planning/ROADMAP.md',
    config_path: '.planning/config.json',
+    skill_paths: skillPaths,
  };


skill_paths is set to a single global directory ($HOME/.claude/skills). The templates also mention a project-local .claude/skills fallback/override and the field name is plural, suggesting multiple lookup locations. Consider outputting an ordered list of skill search paths (project-local first, then home) or renaming the field to skill_path if only one location is intended.

Copilot · 2026-03-02T04:56:44Z

templates/workflows/execute-phase.md

       - .planning/STATE.md (State)
       - .planning/config.json (Config, if exists)
       - ./CLAUDE.md (Project instructions, if exists — follow project-specific guidelines and coding conventions)
-       - .agents/skills/ (Project skills, if exists — list skills, read SKILL.md for each, follow relevant rules during implementation)
+       - ~/.claude/skills/ (Skills, if exists — list skills, read SKILL.md for each, follow relevant rules during implementation)
       </files_to_read>


PR description says skill path references were updated “across executor and orchestrator”, but there are still multiple templates referencing .agents/skills/ (e.g., templates/workflows/plan-phase.md, templates/workflows/quick.md, and several agent templates like maxsim-planner.md). If the skills directory move is intended repo-wide, those remaining references will keep pointing at the old location and create inconsistent behavior/documentation.

Copilot · 2026-03-02T04:56:44Z

templates/workflows/execute-plan.md

+WAVE_FIRST_COMMIT=$(echo "${TASK_COMMITS[0]}" | awk '{print $NF}')
+git diff ${WAVE_FIRST_COMMIT}^..HEAD --name-only


<wave_code_review> says it reviews only the current wave, but WAVE_FIRST_COMMIT is derived from TASK_COMMITS[0]. Since TASK_COMMITS is only ever appended to in this template and never reset per wave, this diff range will start from the very first task of the entire run, not the wave. Consider resetting the commit list at wave start or tracking a per-wave first commit hash explicitly so the review diff is scoped correctly.

Suggested change

WAVE_FIRST_COMMIT=$(echo "${TASK_COMMITS[0]}" | awk '{print $NF}')

git diff ${WAVE_FIRST_COMMIT}^..HEAD --name-only

# WAVE_FIRST_COMMIT must be set to the first commit hash of this wave at wave start, e.g.:

# WAVE_FIRST_COMMIT=$(git rev-parse --short HEAD)

git diff "${WAVE_FIRST_COMMIT}^"..HEAD --name-only

Copilot AI review requested due to automatic review settings March 2, 2026 04:52

Copilot started reviewing on behalf of maystudios March 2, 2026 04:52 View session

Copilot AI reviewed Mar 2, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat(execution): wire Execute→Review→Simplify→Review pipeline into executor workflow#20

feat(execution): wire Execute→Review→Simplify→Review pipeline into executor workflow#20
maystudios wants to merge 1 commit intomainfrom
worktree-agent-a9428419

maystudios commented Mar 2, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Mar 2, 2026

Uh oh!

Copilot AI Mar 2, 2026

Uh oh!

Copilot AI Mar 2, 2026

Uh oh!

Copilot AI Mar 2, 2026

Uh oh!

Copilot AI Mar 2, 2026

Uh oh!

Copilot AI Mar 2, 2026

Uh oh!

Copilot AI Mar 2, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

		# Get all files changed in this wave
		WAVE_FIRST_COMMIT=$(git log --oneline --all --grep="{phase}-{first_plan_in_wave}" --reverse \| head -1 \| cut -d' ' -f1)

	- ~/.claude/skills/ (Skills, if exists — list skills, read SKILL.md for each, follow relevant rules during implementation)
	- ./.claude/skills/ (Project-local skills, if exists — override or extend global skills; list skills and read SKILL.md for each)
	- ~/.claude/skills/ (Global skills, if exists — list skills, read SKILL.md for each, follow relevant rules during implementation)

		WAVE_FIRST_COMMIT=$(echo "${TASK_COMMITS[0]}" \| awk '{print $NF}')
		git diff ${WAVE_FIRST_COMMIT}^..HEAD --name-only

-WAVE_FIRST_COMMIT=$(echo "${TASK_COMMITS[0]}" | awk '{print $NF}')
-git diff ${WAVE_FIRST_COMMIT}^..HEAD --name-only
+# WAVE_FIRST_COMMIT must be set to the first commit hash of this wave at wave start, e.g.:
+#   WAVE_FIRST_COMMIT=$(git rev-parse --short HEAD)
+git diff "${WAVE_FIRST_COMMIT}^"..HEAD --name-only

Conversation

maystudios commented Mar 2, 2026

Summary

Test plan

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Mar 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 2, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants