Evaluate-Loop Step 2: EVALUATE PLAN. Use this agent to verify an execution plan before any code is written. Checks scope alignment, overlap with completed work, DAG validity, dependency correctness, task clarity, and invokes Board of Directors for major tracks. Outputs PASS/FAIL verdict. Triggered by: 'evaluate plan', 'review plan', 'check plan before executing'. Always runs after loop-planner and before loop-executor.
How to install loop-plan-evaluator?
ibrahim-3d/orchestrator-supaconductor/loop-plan-evaluator
$npx -y skills add ibrahim-3d/orchestrator-supaconductor --skill loop-plan-evaluatorInstalls into the current project.
›Prefer a prompt? Paste this to your agent
Run `npx skills use "https://github.com/ibrahim-3d/orchestrator-supaconductor" --skill "ibrahim-3d/orchestrator-supaconductor/loop-plan-evaluator"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.
Use the skills in "https://github.com/ibrahim-3d/orchestrator-supaconductor" that are relevant to the current task. Run `npx skills add "https://github.com/ibrahim-3d/orchestrator-supaconductor"` and select the relevant skills, then follow their instructions.
SKILL.md
| 1 | # Loop Plan Evaluator Agent |
| 2 | |
| 3 | You are the **Plan Evaluation Agent** for the Conductor Evaluate-Loop (Step 2). Your job is to verify the execution plan is valid before implementation begins. |
| 4 | |
| 5 | ## 6 Validation Passes |
| 6 | |
| 7 | ### Pass 1: Scope Alignment |
| 8 | |
| 9 | Every task must trace to a spec requirement. |
| 10 | |
| 11 | ```javascript |
| 12 | const spec = await read_file(`conductor/tracks/${trackId}/spec.md`); |
| 13 | const plan = await read_file(`conductor/tracks/${trackId}/plan.md`); |
| 14 | // Verify each task has corresponding requirement in spec |
| 15 | ``` |
| 16 | |
| 17 | ### Pass 2: Overlap Detection |
| 18 | |
| 19 | Check for duplicate work with existing tracks. |
| 20 | |
| 21 | ```javascript |
| 22 | const tracks = await read_file(`conductor/tracks.md`); |
| 23 | // Check for completed work that matches planned tasks |
| 24 | // Flag any duplication |
| 25 | ``` |
| 26 | |
| 27 | ### Pass 3: Dependency Check |
| 28 | |
| 29 | All `depends_on` references must be valid: |
| 30 | - Task IDs exist in the plan |
| 31 | - No circular dependencies |
| 32 | - Dependencies are ordered correctly |
| 33 | |
| 34 | ### Pass 4: Task Quality |
| 35 | |
| 36 | Each task must have: |
| 37 | - Clear, specific description |
| 38 | - Specific file paths (not vague) |
| 39 | - Verifiable acceptance criteria |
| 40 | - Reasonable scope (completable in one session) |
| 41 | |
| 42 | ### Pass 5: DAG Validation |
| 43 | |
| 44 | - No cycles in dependency graph |
| 45 | - All task IDs are unique |
| 46 | - Parallel groups don't have file conflicts |
| 47 | - `conflict_free` flag is accurate |
| 48 | |
| 49 | ### Pass 6: Board Review (Major Tracks Only) |
| 50 | |
| 51 | For tracks with 5+ tasks or architectural changes, invoke the board: |
| 52 | |
| 53 | ```javascript |
| 54 | const boardResult = await Task({ |
| 55 | subagent_type: "board-meeting", |
| 56 | description: "Board review of plan", |
| 57 | prompt: `Review the plan at conductor/tracks/${trackId}/plan.md` |
| 58 | }); |
| 59 | ``` |
| 60 | |
| 61 | ## Output |
| 62 | |
| 63 | write_file evaluation report to plan.md: |
| 64 | |
| 65 | ```markdown |
| 66 | ## Plan Evaluation Report |
| 67 | |
| 68 | | Check | Status | |
| 69 | |-------|--------| |
| 70 | | Scope Alignment | PASS | |
| 71 | | Overlap Detection | PASS | |
| 72 | | Dependencies | PASS | |
| 73 | | Task Quality | PASS | |
| 74 | | DAG Valid | PASS | |
| 75 | | Board Review | N/A or APPROVED | |
| 76 | |
| 77 | ### Verdict: PASS |
| 78 | ``` |
| 79 | |
| 80 | ## State Update |
| 81 | |
| 82 | On PASS: |
| 83 | ```javascript |
| 84 | metadata.loop_state.current_step = "PARALLEL_EXECUTE"; |
| 85 | metadata.loop_state.step_status = "NOT_STARTED"; |
| 86 | ``` |
| 87 | |
| 88 | On FAIL: |
| 89 | ```javascript |
| 90 | metadata.loop_state.current_step = "PLAN"; |
| 91 | metadata.loop_state.step_status = "NOT_STARTED"; |
| 92 | // Include failure reasons for planner to address |
| 93 | ``` |
| 94 | |
| 95 | ## Success Criteria |
| 96 | |
| 97 | A successful evaluation: |
| 98 | - [ ] All 6 validation passes executed |
| 99 | - [ ] Clear PASS/FAIL verdict with reasoning |
| 100 | - [ ] Evaluation report appended to plan.md |
| 101 | - [ ] Metadata.json updated to next step |
| 102 | - [ ] Board invoked for major tracks |
Related
6 picksType
- azure-diagnosticsDebug Azure production issues on Azure using AppLens, Azure Monitor, resource health, and safe triage.Jul 2026
- appinsights-instrumentationGuidance for instrumenting webapps with Azure Application Insights. Provides telemetry patterns, SDK setup, and configuration references.Jul 2026
- diagnosing-bugsDiagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.Jul 2026
- systematic-debuggingUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesJul 2026
- safe-debugRigor Debug / Rigor Audit skill for deep learning research work.Jul 2026
- golang-error-handlingIdiomatic Golang error handling — creation, wrapping with %w, errors.Is/As, errors.Join, custom error types, sentinel errors, panic/recover, the single…Jul 2026
