skills.rest
l0lxl0lw
l0lxl0lwCOMMUNITY
Explore all skills by this author →
GitHub

omc-plan

Generates structured work plans through interview, direct, consensus, and review modes.

Updated Dec 18, 2025
Search Tags:#planning#software design#adr#interview#consensus#requirements gathering#plan review
One-click install
npx skills add https://github.com/l0lxl0lw/dotfiles --skill omc-plan-l0lxl0lw
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: omc-plan
Source: https://github.com/l0lxl0lw/dotfiles/tree/main/ai/shared/skills/omc/plan
Command: npx skills add https://github.com/l0lxl0lw/dotfiles --skill omc-plan-l0lxl0lw

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Jumping into code without clear requirements causes rework, scope creep, and missed edge cases. This Skill turns vague or detailed requests into quality-gated, actionable work plans before any implementation begins. ## Core Features & Use Cases - Adaptive Interview Mode: Detects broad requests and gathers requirements one question at a time, using an explore agent to answer codebase questions before asking the user. - Consensus Mode (RALPLAN-DR): Runs a sequential Planner, Architect, and Critic loop with structured deliberation, producing an ADR with decision drivers, alternatives, and consequences. - Review Mode: Evaluates existing plans from .omc/plans/ via a Critic agent and returns APPROVED, REVISE, or REJECT verdicts. - Use Case: A user says "plan the migration of our auth system to OAuth2". The Skill interviews for scope, consults an Analyst for risks, runs the consensus loop in deliberate mode due to the security-sensitive domain, and saves an approved plan with testable acceptance criteria to .omc/plans/. ## Quick Start Ask the assistant to plan your next feature, for example: "Plan the implementation of rate limiting for our public API using the omc-plan skill."

Frequently Asked Questions about omc-plan

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

FAQPage Schema
How do I create a work plan before writing code?▼

Invoke the plan workflow with a request like "plan this feature". Broad requests trigger an interactive interview that asks one question at a time, while detailed requests generate a plan directly. The finished plan is saved to .omc/plans/.

What is consensus mode in plan generation?▼

Consensus mode runs a sequential Planner, Architect, and Critic loop using RALPLAN-DR structured deliberation until the Critic approves or five iterations pass. The final output includes an ADR with decision, drivers, alternatives, and consequences.

How do I review an existing implementation plan?▼

Use the --review flag or say "review this plan". The Skill reads the plan from .omc/plans/ and has a Critic agent return an APPROVED, REVISE with feedback, or REJECT verdict.

When should I use deliberate mode versus short mode?▼

Use deliberate mode via --deliberate for high-risk work such as auth, security, data migrations, destructive changes, or compliance-sensitive features. It adds a pre-mortem with three failure scenarios and an expanded unit, integration, e2e, and observability test plan.

When should I skip planning and just execute?▼

Skip planning for single focused fixes with obvious scope or when the user says "just do it". In those cases the workflow hands off directly to the ralph execution skill instead of producing a plan.

Related Skills

Explore More Software Engineering Skills→
HACK-WUHACK-WU
2

code-survey

Surveys codebase dimensions relevant to a requirement before design work begins.

Community
Advanced
gautam-achieveaigautam-achieveai
2

implement

Implement approved plans into code using TDD and verification.

Community
Advanced
alatyshaualatyshau

loki

Challenges plans and assumptions through adversarial devil's advocate analysis.

Community
Basic
SensLiaoSensLiao

gsd-fast

Execute one-sentence tasks directly in the current context without subagents.

Community
Basic