How do I install the Evolve Plan Review skill?

Run npx skillmds@latest add aibot88/evolve-plan-review in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Evolve Plan Review safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Evolve Plan Review?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Evolve Plan Review free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Evolve Plan Review?

aibot88 (@aibot88) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Evolve Plan Review by aibot88 — Agent Skill

Name: Evolve Plan Review
Rating: 4.666666666666667 (3 reviews)
Author: aibot88

Evolve Plan Review

Use when scout-report.md exists and TDD/Build hasn't started yet. Runs four lenses (CEO, Eng, Design, Security) in parallel on the task list and produces a verdict (PROCEED, REVISE, ABORT) before code is written. Catches misaligned plans before they cost cycles.

aibot88 Updated Sep 9, 2026 3 repo stars

File contents

evolve-plan-review

Sprint 2 multi-lens review (v8.16+). Inspired by garrytan/gstack/autoplan. Default-off via EVOLVE_PLAN_REVIEW=0; enable per cycle.

When to invoke

After Scout writes <workspace>/scout-report.md, before TDD/Build
Whenever the cycle goal mentions architecture, sandbox, or kernel changes
When the previous cycle aborted due to misaligned scope

When NOT to invoke

Cycle is a documented retry of a previously-approved plan
Single-line fix or trivial bug
Eval-only cycles (no code change)

Workflow

Step	Action	Exit criteria
1	Verify `<workspace>/scout-report.md` exists + fresh	scout-report.md valid
2	Dispatch 4 lenses in parallel via `subagent-run.sh dispatch-parallel plan-reviewer`	4 worker artifacts
3	Aggregator computes verdict (PROCEED/REVISE/ABORT)	`<workspace>/plan-review.md` present, first line is `Verdict: <X>`
4	Phase gate `gate_plan_review_to_tdd` enforces verdict	Gate passes only on PROCEED

Verdict semantics

Verdict	Trigger	Orchestrator action
`PROCEED`	Avg score ≥ 7 AND no lens < 5	Advance to TDD
`REVISE`	Avg ≥ 5 AND any lens < 5	Re-run Scout (max 2 retries)
`ABORT`	Any lens explicit ABORT, OR avg < 5	End cycle

Output contract

<workspace>/plan-review.md with first line Verdict: <X>, second Average Score: <N.N>, then per-lens reports.

Composition

Invoked by:

/plan-review slash command
evolve-loop macro (between Scout and TDD when EVOLVE_PLAN_REVIEW=1)

The plan-reviewer persona uses parallel_subtasks (see .evolve/profiles/plan-reviewer.json) — four lens sub-personas run concurrently and merge via aggregator.sh phase=plan-review.

Reference

.evolve/profiles/plan-reviewer.json for lens prompt templates
scripts/aggregator.sh plan-review merge mode
scripts/phase-gate.sh:gate_plan_review_to_tdd
docs/architecture/tri-layer.md (anti-patterns)

---name: evolve-plan-reviewdescription: Use when scout-report.md exists and TDD/Build hasn't started yet. Runs four lenses (CEO, Eng, Design, Security) in parallel on the task list and produces a verdict (PROCEED, REVISE, ABORT) before code is written. Catches misaligned plans before they cost cycles.--- # evolve-plan-review > Sprint 2 multi-lens review (v8.16+). Inspired by `garrytan/gstack/autoplan`. Default-off via `EVOLVE_PLAN_REVIEW=0`; enable per cycle. ## When to invoke - After Scout writes `<workspace>/scout-report.md`, before TDD/Build - Whenever the cycle goal mentions architecture, sandbox, or kernel changes - When the previous cycle aborted due to misaligned scope ## When NOT to invoke - Cycle is a documented retry of a previously-approved plan - Single-line fix or trivial bug - Eval-only cycles (no code change) ## Workflow | Step | Action | Exit criteria | |---|---|---| | 1 | Verify `<workspace>/scout-report.md` exists + fresh | scout-report.md valid | | 2 | Dispatch 4 lenses in parallel via `subagent-run.sh dispatch-parallel plan-reviewer` | 4 worker artifacts | | 3 | Aggregator computes verdict (PROCEED/REVISE/ABORT) | `<workspace>/plan-review.md` present, first line is `Verdict: <X>` | | 4 | Phase gate `gate_plan_review_to_tdd` enforces verdict | Gate passes only on PROCEED | ## Verdict semantics | Verdict | Trigger | Orchestrator action | |---|---|---| | `PROCEED` | Avg score ≥ 7 AND no lens < 5 | Advance to TDD | | `REVISE` | Avg ≥ 5 AND any lens < 5 | Re-run Scout (max 2 retries) | | `ABORT` | Any lens explicit ABORT, OR avg < 5 | End cycle | ## Output contract `<workspace>/plan-review.md` with first line `Verdict: <X>`, second `Average Score: <N.N>`, then per-lens reports. ## Composition Invoked by: - `/plan-review` slash command - `evolve-loop` macro (between Scout and TDD when `EVOLVE_PLAN_REVIEW=1`) The `plan-reviewer` persona uses `parallel_subtasks` (see `.evolve/profiles/plan-reviewer.json`) — four lens sub-personas run concurrently and merge via `aggregator.sh phase=plan-review`. ## Reference - `.evolve/profiles/plan-reviewer.json` for lens prompt templates - `scripts/aggregator.sh` plan-review merge mode - `scripts/phase-gate.sh:gate_plan_review_to_tdd` - `docs/architecture/tri-layer.md` (anti-patterns)

Frequently asked questions