Skip to content
Agentmart
Submit a tool

ralplan

Consensus planning via Planner + Architect + Critic triad before execution

Skill by MaxFerAlten

IndexedMIT
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/maxferalten-ralplan

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Consensus planning via Planner + Architect + Critic triad before execution

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/ralplan

What's inside

Skill name
ralplan
Description
Consensus planning via Planner + Architect + Critic triad before execution
License
MIT
Instructions
100 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# /ralplan — Consensus Planning

## Trigger

`/ralplan <task>` or "ralplan", "consensus plan", "planner architect critic", "planning triad"

## Description

Ralplan runs a **three-role planning triad** — Planner, Architect, and Critic — that must reach consensus before any code is written. It prevents premature execution by requiring all three roles to converge on a single implementation plan.

## When to Use

- Feature is large enough to benefit from multi-perspective planning
- Previous attempts at the task produced conflicting approaches
- Stakeholder (user) wants auditability before execution starts
- Coming from `deep-interview` or `deepsearch` with a crystallized spec

## Do Not Use When

- Task is a trivial one-liner fix
- User explicitly says "just do it" or invokes `/autopilot` directly
- A consensus plan already exists in `.omx/plans/`

## Roles

| Role | Responsibility |
|------|----------------|
| **Planner** | Proposes the implementation approach: files, steps, interfaces |
| **Architect** | Evaluates structural fit, technical debt, scalability, coupling |
| **Critic** | Stress-tests assumptions, identifies risks, edge cases, missing tests |

## Flow

### Phase 0: Context Intake
1. Load any existing spec from `.omx/specs/` or `.tenderclaw/state/plans/`
2. Summarize the task in ≤ 3 sentences

### Phase 1: Planner Turn
1. Produce a concrete implementation plan:
   - Files to create or modify (exact paths)
   - Key functions / classes to add
   - Step order with rationale
2. Estimate complexity: Low / Medium / High
3. Flag any external dependencies

### Phase 2: Architect Turn
1. Review the Planner's proposal
2. Evaluate: does it fit the existing architecture?
3. Propose structural improvements or alternatives
4. Rate feasibility: 1–5

### Phase 3: Critic Turn
1. Review both Planner and Architect outputs
2. Identify: missing tests, security gaps, performance risks, rollback difficulty
3. Assign risk level: Low / Medium / High / Critical
4. Propose mitigations for each risk

### Phase 4: Consensus Round
Repeat (up to `--rounds` times, default 2) until consensus:
- Each role re-evaluates after seeing the others' input
- Consensus = all three rates feasibility >= 3 AND risk <= Medium
- If no consensus after max rounds: output divergence report and halt

### Phase 5: Crystallize Plan
Write final consensus plan to `.omx/plans/ralplan-{slug}-{timestamp}.md`:
```
# Consensus Plan: {task}
## Approach (Planner)
## Architecture Notes (Architect)
## Risk Register (Critic)
## Final Steps (ordered)
## Acceptance Criteria
```

## Execution Bridge

After crystallization, offer:
1. **$autopilot** — Execute the consensus plan (recommended)
2. **$ralph** — Persistent execution with re-verification
3. **$team** — Parallel worker execution
4. **Refine further** — Another consensus round

## Rules

- Never emit code during planning phases
- Planner proposes, Architect validates, Critic challenges — roles do not swap
- If Critic rates risk Critical, HALT and request user guidance
- Consensus plan is the single source of truth for execution

## Agents

- metis
- sisyphus
- sentinel

## Keywords

- ralplan, consensus plan, planner architect critic, planning triad, three-role planning

From MaxFerAlten/tenderclaw, MIT.

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Apr 17, 2026 (6 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from MaxFerAlten

Cite this page

ralplan skill. Agent Mart. https://agentmart.dev/skills/maxferalten-ralplan (last updated 2026-10-24)