notque/vexjoy-agent / skills/process/workflow
workflow
VerifiedStructured work: multi-phase tasks, feature builds, planning, objective loops, hill climbing.
references/pl-check.md
View on GitHub# Plan Checker Reference
## Overview
Validate plans before execution using goal-backward analysis. Start from the stated goal and verify every requirement has a complete path through the plan to completion. This catches plan-level defects before they waste an entire execution cycle.
**Key principle**: Plan completeness does not equal goal achievement. A plan can have all tasks filled in, each well-specified, and still miss the goal entirely. The checker validates by working backward from the goal through every requirement to verify complete coverage -- not just that each task looks reasonable in isolation.
This reference operates as a gate between planning and execution, validating across 10 dimensions and issuing a PASS/BLOCK verdict. If issues are found, a bounded revision loop allows up to 3 iterations before proceeding with documented risks (because the cost of continued planning is not zero -- it consumes context, delays execution, and each revision may introduce new issues).
## Instructions
### Phase 1: LOAD
Load the plan and its context. The checker needs three things: the plan itself, the goal it serves, and the repository rules it must comply with.
**Step 1: Identify the plan**
Accept the plan from one of these sources (in priority order):
1. **Explicit path**: User provides a file path to the plan
2. **Feature state**: Check `.feature/state/plan/` for the active feature plan
3. **Active plans**: Check `plan/active/` for workflow-orchestrator plans
4. **Inline plan**: User pastes plan content directly
If no plan is found, BLOCK with a message telling the user to provide a plan path or run `/feature-lifecycle` first.
**Step 2: Extract the goal**
The goal is the anchor for goal-backward analysis. Find it in the plan's own `## Goal` or `## Success Criteria` section, the parent design document (for feature plans: `.feature/state/design/`), or the user's original request (for workflow-orchestrator plans).
If no goal is found, BLOCK: goal-backward analysis requires a goal. Direct the user to add a `## Goal` section.
**Step 3: Load repository context**
Read the target repository's CLAUDE.md (if present) for dimension 9 validation:
```bash
find . -maxdepth 2 -name "CLAUDE.md" -type f 2>/dev/null | head -5
```
**Step 4: Extract requirements**
From the goal and design document, extract a numbered list of requirements. Each becomes a row in the coverage matrix. Success criteria in the plan count as requirements.
**GATE**: Plan loaded. Goal identified. Requirements extracted. Proceed to CHECK.
---
### Phase 2: CHECK
Run goal-backward analysis across all 10 verification dimensions. For each, produce structured findings or mark as PASS.
Full dimension specifications (severity, checks, tables, examples) for all 10 dimensions — Requirement Coverage, Task Completeness, Dependency Correctness, Key Links Planned, Scope Sanity, Verification Derivation, Context Compliance, Cross-Plan Data Contracts, CLAUDE.md Compliance, Achievability — live in `check-dimensions.md` (ported from the legacy `plan-checker/references/dimensions.md`).
**Dimension severity summary**:
| # | Dimension | Severity |
|---|-----------|----------|
| 1 | Requirement Coverage | Blocker |
| 2 | Task Completeness | Warning |
| 3 | Dependency Correctness | Blocker |
| 4 | Key Links Planned | Blocker |
| 5 | Scope Sanity | Warning at 4, Blocker at 5+ |
| 6 | Verification Derivation | Warning |
| 7 | Context Compliance | Blocker |
| 8 | Cross-Plan Data Contracts | Blocker |
| 9 | CLAUDE.md Compliance | Blocker |
| 10 | Achievability | Warning |
**GATE**: All 10 dimensions checked. Findings collected. Proceed to VERDICT.
---
### Phase 3: VERDICT
Compile findings and issue verdict.
**Step 1: Compile findings** — Order by severity (blockers first), then by dimension number. Each finding uses the structured format (Plan, Dimension, Severity, Description, Fix hint) — see `check-verdict-format.md` (ported from legacy `plan-checker/references/verdict-format.md`).
**Step 2: Issue verdict**
| Condition | Verdict | Action |
|-----------|---------|--------|
| Zero findings | **PASS** | Proceed to execution |
| Warnings only, no blockers | **PASS with warnings** | Proceed, but address warnings if time allows |
| Any blocker findings | **BLOCK** | Plan must be revised before execution |
**Step 3: Format output** — Full verdict output template (headers, findings block, coverage matrix block) lives in `check-verdict-format.md`.
If verdict is PASS or PASS with warnings, direct the user to proceed with `/feature-lifecycle` (implement phase) or workflow-orchestrator EXECUTE.
If verdict is BLOCK, proceed to Phase 4 (Revision Loop).
**GATE**: Verdict issued. If PASS, done. If BLOCK, proceed to revision loop.
---
### Phase 4: REVISION LOOP (only if BLOCK)
Bounded revision loop: fix blocker findings, re-check, max 3 iterations. After 3 good-faith attempts, remaining issues are either genuinely hard (better discovered during execution with real code) or low-probability (not worth further planning time).
**Step 1: Revise plan** — For each blocker, apply the fix_hint. If auto-revise is OFF (default), present to the user for approval; if ON, apply directly.
**Step 2: Re-check** — Run Phase 2 again on the revised plan. Only re-check dimensions that had findings (dimensions that passed don't regress from plan edits).
**Step 3: Evaluate**
| Condition | Action |
|-----------|--------|
| All blockers resolved | Issue PASS verdict, done |
| Blockers remain, iterations < 3 | Next iteration |
| Blockers remain, iterations = 3 | Document remaining as known risks, issue **PASS (with known risks)** |
Revision loop output template (iteration tracking, max-iterations banner, known-risks section): `check-verdict-format.md`.
---
## Error Handling
Common errors (no plan found, no goal found, file verification fails, CLAUDE.md not found, revision loop exhausted, plan is inline text) and solutions: `check-errors.md` (ported from legacy `plan-checker/references/errors.md`).
## References
- ADR-074: Plan Checker Pre-Execution Validation (historical reference -- file not present in current repo)
- [Feature Lifecycle](../../feature-lifecycle/SKILL.md) -- plan phase produces plans this reference validates; implement phase executes plans after validation
- [Workflow Orchestrator](../../workflow/references/workflow-orchestrator.md) -- PLAN phase produces plans this reference can validate
- [Verification Before Completion](../../verification-before-completion/SKILL.md) -- post-execution counterpart (validates results, not plans)
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| planning-and-task-breakdownaddyosmani/agent-skillsBreaks work into ordered tasks. Use when you have a spec or clear requirements and need to break work into implementable tasks. Use when a task feels too large to start, when you need to estimate scope, or when parallel work is possible. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| screenpipethedotmack/claude-memSearch screen and audio history recorded by Screenpipe and optionally save selected context into claude-mem. Use when the user asks to recall desktop activity, a meeting, or something seen in another app through Screenpipe. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| contact-widgetnexu-io/open-designSelf-contained floating chat widget with welcome screen, social links, meeting button, and message input. Single HTML file, zero dependencies. | Stars: 108k | Updated | Claude CodeCodex | Verified | Install |
| meeting-notesnexu-io/open-designMeeting notes page — title bar with attendees, agenda checklist, decisions block, action items table with owners + dates, and a "next meeting" footer. Use when the brief mentions "meeting notes", "minutes", "1:1 notes", "all-hands recap", or "会议纪要". | Stars: 108k | Updated | Claude CodeCodex | Verified | Install |