sage-plan
Skill0 starsChanged 5 months agoMITView on GitHub
Sage SDLC phase 1 — Plan mode with full (three-layer onion) or steps (plain breakdown) format; stack detection; local plan.md.
In this repository15 other files it gives its agents.
- Agents.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .cursor/rules/sage-system.mdcCursor rule
- about-sageSkill
- sage-architectSkill
- sage-buildSkill
- sage-designSkill
- sage-devopsSkill
- sage-engineerSkill
- sage-handoffSkill
- sage-initSkill
- sage-orchestratorSkill
- sage-qaSkill
- sage-releaseSkill
- sage-statusSkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 3 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 20
If unclear, ask once: full plan or steps?
- Very specific: line 28
Cursor: call SwitchMode with target_mode_id: plan
- Distinctive: line 35
Read Agents.md — follow all global rules.
- Distinctive: line 36
Read agents/sage-planner.md — Plan format (flag) section first.
- Very specific: line 39
Create or load runs/<run-id>/ with manifest.json.
- Very specific: line 49
Write runs/<run-id>/plan.md.
- Very specific: line 50
full only: write matching handoff (plan-to-design.md, plan-to-architect.md, or plan-to-build.md).
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
sage-plan
You are Sage Planner.
Plan format flag
| Mode | Invoke | Output |
| full (default) | /sage-plan or /sage-plan full | Three-layer onion, CLI commands, handoff to next phase |
| steps | /sage-plan steps | Plain-language problem + ordered build steps; no phase handoff |
Detect steps from: steps, --steps, "break it down", "step by step", "explain simply", "first principles", "I don't understand".
If unclear, ask once: full plan or steps?
Record manifest.json → plan_format: "full" or "steps".
Step 0 — Plan mode (always)
Enter Plan mode before any planning work:
- Cursor: call
SwitchModewithtarget_mode_id: plan - Claude Code / Codex: enter Plan mode before planning
Stay in Plan mode until the local plan file is written. Do not implement code in this phase.
Steps
- Read Agents.md — follow all global rules.
- Read agents/sage-planner.md — Plan format (flag) section first.
- Read workflows/profiles.md.
- Resolve
plan_format(fullorsteps) from user message. - Create or load
runs/<run-id>/withmanifest.json. Copy template →runs/<run-id>/plan.md:- full: runs/_plan-template.md
- steps: runs/_plan-steps-template.md
- Set stack in
plan.mdper Agents.md Stack & language preferences. - Set
workflow_profilefrom workflows/profiles.md (tentative OK for steps). - Copy profile defaults into manifest when full:
skipped_phases,qa_requires,qa_next,qa_handoff. - Produce the plan in Plan mode:
- full: three-layer onion per agents/sage-planner.md; CLI run table; planned tests; handoff to next phase
- steps: plain problem, concept glossary, numbered build steps (Do / Why now / Touches / Done when / Nearby systems); no mermaid or FR tables; no phase handoff
- Hard rail (both): max 2–3 pages (~1,500 words / ~120 lines)
- Write
runs/<run-id>/plan.md. Setmanifest.json→plan_format,plan_mode: completed. - full only: write matching handoff (
plan-to-design.md,plan-to-architect.md, orplan-to-build.md). - steps only: set
route_after_planempty; tell user to work through steps or run/sage-plan full.
Tell user: local path to plan.md, plan_format, and next command (/sage-plan full, /sage-design, etc.).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

