Help us improve
Share bugs, ideas, or general feedback.
From agentops
Creates self-contained, dependency-aware implementation plans for software projects. Use when starting a new project, refining architecture, or decomposing complex work into actionable beads.
npx claudepluginhub boshu2/agentops --plugin agentopsHow this skill is triggered — by the user, by Claude, or both
Slash command
/agentops:planning-workflowThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
<!-- TOC: Philosophy | THE EXACT PROMPT | Process Overview | References -->
Creates structured plans for multi-step tasks including software features, implementations, research, or projects. Deepens plans via interactive sub-agent reviews.
Generates atomic PLAN.md files for hierarchical project planning in solo agentic dev with Claude. Covers briefs, roadmaps, phases; includes tasks, verification, checkpoints, success criteria.
Creates detailed implementation plans through interactive, iterative collaboration. Useful for planning features, refactoring, or tasks.
Share bugs, ideas, or general feedback.
Core Philosophy: "Planning tokens are a lot fewer and cheaper than implementation tokens."
The models are far smarter when reasoning about a detailed plan that fits within their context window. This is the key insight behind spending 80%+ of time on planning.
You have a usable plan when all of the following hold:
br ready --json without re-reading the plan.You have NOT delivered if any of these is true: the plan is < ~1,500 lines for a non-trivial project (under-specified); the plan is > ~10,000 lines with no decomposition (impossible to act on); the human is still being asked "what should this do?" mid-implementation (planning was abandoned, not completed); beads exist but have no dependency edges (the plan's structure was lost in conversion).
Reach for something else if:
codebase-archaeology or idea-wizard first; planning is for execution, not discovery.porting-to-rust or testing-conformance-harnesses; the plan is the spec itself.When a planning model proposes architecture, the proposal is a hypothesis. Ground every load-bearing claim in a verifiable source before letting it survive a review round:
git log, or use codebase-archaeology to confirm structural claims before they shape the plan.<tool> for X," confirm the tool's current contract supports X. Linkrot in plans is a quiet failure mode.A plan that survived review without grounding is a plan that will surprise you in implementation. Cheap verification at planning time beats expensive rework after the code is half-written.
After each review round, before sending the plan back for another pass, run all four:
If any of these fails, the plan is not ready for beads conversion — do another review round.
Carefully review this entire plan for me and come up with your best revisions in terms of better architecture, new features, changed features, etc. to make it better, more robust/reliable, more performant, more compelling/useful, etc. For each proposed change, give me your detailed analysis and rationale/justification for why it would make the project better along with the git-diff style change versus the original plan shown below:
<PASTE YOUR EXISTING COMPLETE PLAN HERE>
After GPT Pro finishes (may take 20-30 minutes), paste output into Claude Code:
OK, now integrate these revisions to the markdown plan in-place; use ultrathink and be meticulous. At the end, you can tell me which changes you wholeheartedly agree with, which you somewhat agree with, and which you disagree with:
```[Pasted text from GPT Pro]```
1. INITIAL PLAN (GPT Pro / Opus 4.7 in web app)
└─► Explain goals, intent, workflows, tech stack
2. ITERATIVE REFINEMENT (GPT Pro Extended Reasoning)
└─► 4-5 rounds of revision until steady-state
3. MULTI-MODEL BLENDING (Optional but recommended)
└─► Gemini 3.1 Pro Deep Think, Grok4 Heavy, Opus 4.7
└─► GPT Pro as final arbiter
4. CONVERT TO BEADS (Claude Code + Opus 4.7)
└─► Self-contained tasks with dependency structure
5. POLISH BEADS (6+ rounds until steady-state)
└─► Cross-model review, never oversimplify
| Good Plan | Great Plan |
|---|---|
| Describes what to build | Explains WHY you're building it |
| Lists features | Details user workflows and interactions |
| Mentions tech stack | Justifies tech choices with tradeoffs |
| Has tasks | Has tasks with dependencies and rationale |
| ~500 lines | ~3,500+ lines after refinement |
| Topic | Reference |
|---|---|
| All exact prompts | PROMPTS.md |
| Real-world examples | EXAMPLES.md |
| FAQ | FAQ.md |