skills.rest
apireno
apirenoCOMMUNITY
Explore all skills by this author →
GitHub

vp-review

Runs multi-persona VP reviews of artifacts and synthesizes a CTO verdict.

1|Updated Jun 13, 2026
Search Tags:#code review#adr#workflow orchestration#multi-agent#openrouter#verdict synthesis#sprint plan
One-click install
npx skills add https://github.com/apireno/agent-workflow-template --skill vp-review-apireno
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: vp-review
Source: https://github.com/apireno/agent-workflow-template/tree/main/.claude/skills/vp-review
Command: npx skills add https://github.com/apireno/agent-workflow-template --skill vp-review-apireno

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Getting consistent, multi-perspective review of engineering artifacts (sprint plans, ADRs, RCAs, PRDs, dev-reports) is hard with a single AI pass — one reviewer misses domain-specific risks. This Skill fans an artifact out to parallel VP personas (product, engineering, security, DevOps, data science) and synthesizes their structured verdicts into one CTO decision. ## Core Features & Use Cases - Parallel VP reviews: Runs vp-prod and vp-eng by default, with specialty VPs (vp-security, vp-devops, vp-datascience) added via the --vps flag when the artifact touches their domain. - Configurable review engines: Choose subagent (in-session, default), kimi (cross-model-family independent reviewer via OpenRouter), codex, gemini, handoff, or claude-p, resolved by flag, environment variable, or repo config file. - Structured synthesis: Tallies BLOCKER/MAJOR/MINOR/NOTE findings, detects cross-VP convergence, and writes an APPROVED / REJECTED-FOR-REVISION / ESCALATE-TO-CEO decision file next to the artifact. - Use Case: After drafting a sprint plan that introduces new CI pipelines, run a review with vp-prod, vp-eng, and vp-devops to get a consolidated verdict with concrete blockers before execution begins. ## Quick Start Ask the agent to review the artifact, for example: run a VP review on docs/plans/sprint-12.md with vp-prod, vp-eng, and vp-devops.

Frequently Asked Questions about vp-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a multi-VP review on a sprint plan or ADR?▼

Invoke the skill with the artifact path, for example /vp-review docs/plans/sprint-12.md. It defaults to vp-prod and vp-eng; add specialty reviewers with --vps=vp-security,vp-devops,vp-datascience or use --vps=all for all five.

Which review engine should I use for AI-generated code review?▼

Use the kimi engine when independence matters, such as reviewing Claude-authored work, because it is a different model family and avoids shared-method bias. The default subagent engine runs in-session at no incremental cost; codex is untested and claude-p is metered break-glass only.

How do I change the default review engine?▼

Pass --engine with the invocation, set the REVIEW_ENGINE environment variable, or write the engine name to a .review-engine file in the repo. The resolve-review-engine.sh script applies that precedence order, defaulting to subagent.

Why does vp-review fail with a CTO home anchoring error?▼

The skill requires a directory containing .cto/projects.yaml and rejects the unconfigured template's placeholder registry. Run it from the CTO home or set a permanent anchor with: echo /path/to/<project>-cto > ~/.cto/home.

What output does a VP review produce?▼

Each VP writes a structured verdict with a recommendation and findings by severity to a temp directory. The CTO then writes a cto-decision-<timestamp>.md file beside the artifact containing the final decision, top blockers, convergence findings, and action items.

Related Skills

Explore More Software Engineering Skills→
LeGeRyChEeSeLeGeRyChEeSe

bmad-create-story

Generates comprehensive developer-ready story files from epics, architecture, and sprint status.

Community
Advanced
jguadagnojguadagno
18

reviewer-protocol

Enforce author lockout policies during reviewer rejection workflows.

Community
Advanced
apexradiusapexradius

adversarial-verify

Verify deliverable done-claims with evidence-backed adversarial checks.

Official
Advanced
robertsthomasrobertsthomas

critique

Evaluate UX design quality with Nielsen heuristics and cognitive-load scoring.

Community
Advanced