lulu-workflow
Skill3 starsChanged 6 days agoMITView on GitHub
Top-level development workflow framework. Use when mentioning lulu-workflow, 开发工作流, dev workflow, product doc workflow, 产品文档流程, lulu-brainstorm, or any sub-stage (lulu-bet, lulu-approach, lulu-spec, lulu-design, lulu-plan, lulu-tasks, lulu-exec).
In this repository11 other files it gives its agents.
- composeSkill
- decisionSkill
- eval-librarySkill
- lulu-approachSkill
- lulu-betSkill
- lulu-brainstormSkill
- lulu-designSkill
- lulu-execSkill
- lulu-planSkill
- lulu-specSkill
- lulu-tasksSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 0 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 25
Read the key for the current $CYCLE_TYPE: each entry maps a from stage to the to stages it allows, and a null from lists the stages a cycle …
- Very specific: line 61
Call $CYCLE_CONTROL only. A bare python3 path or cycle_schema.py is outside this file.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
lulu-workflow
Routes project-level work into a stage skill. Done when init has reported, or the matching sub-SKILL is loaded.
Runtime modules (loaded by sub-skills, not this file):
_runtime.md— Script Macros + Platform / Session / Execution Mode (all sub-skills)_slowpath.md— Feature Resolution Slow Path (loaded on demand)_transitions.md— Stage Transitions + Rollback (loaded at delivery)_subagent.md— Sub-agent Context (lulu-exec, lulu-tasks, decision)
Stage lines
$CYCLE_TYPE is topic or feature, and each has its own stage dependencies in config/transition-table.json, under the topic and feature keys.
Read the key for the current $CYCLE_TYPE: each entry maps a from stage to the to stages it allows, and a null from lists the stages a cycle can start at.
Routing
Load one sub-SKILL after $CYCLE_ID is confirmed.
| Key | Sub-SKILL | Action |
lulu-brainstorm | Divergence before a problem is defined | Read lulu-brainstorm/SKILL.md |
lulu-bet / pd | Product decision | Read lulu-bet/SKILL.md |
lulu-spec / ps | Product spec (feature) | Read lulu-spec/SKILL.md |
lulu-approach / td | Tech decision | Read lulu-approach/SKILL.md |
lulu-design / ds | Tech design | Read lulu-design/SKILL.md |
lulu-plan / t | Tech plan | Read lulu-plan/SKILL.md |
lulu-tasks / w | Tech work order | Read lulu-tasks/SKILL.md |
lulu-exec / c | Task exec | Read lulu-exec/SKILL.md |
- The sub-SKILL owns drafting, evaluation, and delivery.
Bootstrap
First use in a repo. Read _runtime.md § Script Macros and § Platform Context before the command.
- When — the repo has no platform hooks, or this is the first run.
- Run —
$CYCLE_CONTROL init-project - Done — report success or stderr. Re-run is safe.
- Next — enter a sub-SKILL. Cycle binding is
_runtime.md§ Session Foundation.
Script Macros
$CYCLE_CONTROL is defined in _runtime.md § Script Macros. This file invokes init-project only.
start,menu, andresolve-tokenrun from_slowpath.md.bind-contextruns from_runtime.md§ Session Foundation after$CYCLE_IDis confirmed.- Subcommands and stdout:
cycle_control.pymodule docstring or--help. - Non-zero exit → stop and report stderr.
- Call
$CYCLE_CONTROLonly. A barepython3path orcycle_schema.pyis outside this file.
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.

