intake-engine
Skill4 starsChanged 8 months agoMITView on GitHub
StackMoss CLI intake flow — Fast mode (6 questions), Interview mode (12 questions), persona detection, decision tree, pack overlay composition, and completeness gates. Use when working on the onboarding questionnaire or pack selection logic.
In this repository6 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- cli-pipelineSkill
- stretch-testingSkill
- template-engineSkill
- skill-creatorSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 26
Completeness gate: Only check Q5. If missing → re-ask Q5 once.
- Distinctive: line 69
Never infer, never assume, never use default for critical fields
- Distinctive: line 78
Use LLM for pack/role selection logic
- Distinctive: line 80
Add questions not in the BRD schema
≈ “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
Intake Engine Skill
When to Use
- Implementing the intake questionnaire (Fast/Interview mode)
- Building the persona detection + pack selection logic
- Working on completeness gates
- Implementing the 2D Matrix overlay compose (UserType × ProjectType)
Intake Modes
Fast Mode (6 questions, ~3 min)
| # | Question | Choices |
| Q1 | Role in project? | Biz lead / Dev lead / Solo / Unknown |
| Q2 | Who is the product for? | Individual / SME / Enterprise / Community |
| Q3 | Geography? | Internal / VN / Global |
| Q4 | Primary channel? | Web / Mobile / Chat / API |
| Q5 | Sensitive data? | None / PII / Finance / Compliance |
| Q6 | First feature to ship? | (free text + appetite S/M/L) |
Completeness gate: Only check Q5. If missing → re-ask Q5 once. Do NOT check outcome/non-goals (avoid turning Fast into Interview).
Interview Mode (12 questions, ~10 min)
- Block 1 (Q1–Q4): Business context
- Block 2 (Q5–Q8): Technical constraints
- Block 3 (Q9–Q12): Team & velocity
Completeness gate: Full check — must have "for who", measurable outcome, at least 1 non-goal. If missing → up to 2 targeted follow-up questions.
2D Matrix: Pack Selection
UserType overlay × ProjectType overlay → Role set| UserType | ProjectType | Roles |
| BizLed | MVP | TL, BA, DEV, QA(light), DOCS |
| BizLed | Production | TL, BA, DEV, QA(strong), OPS(light), DOCS |
| DevLed | MVP | TL, DEV, QA, DOCS (BA-lite in TL) |
| DevLed | Production | TL, DEV, QA(strong), OPS, DOCS |
| Solo | MVP | TL, DEV, QA(light) |
| Solo | Production | TL, DEV, QA, DOCS |
| Newbie | any | TL(guide), DEV(small), QA(checklist), DOCS |
Auto-add Rules (no user prompt needed)
- Q5/Q6 = PII / Finance / Compliance → add SEC-lite
- Q7 = VPS / Cloud AND Q2 = SME / Enterprise → add OPS-lite
Architecture Rules
Decision Tree (deterministic, no LLM)
src/intake/
├── questions.ts → Question definitions + validation
├── fast-mode.ts → 6-question flow + completeness gate
├── interview-mode.ts → 12-question flow + completeness gate
├── pack-selector.ts → 2D matrix overlay compose
└── auto-add.ts → SEC-lite / OPS-lite auto-detectionSkipped Questions
- Any skipped question → write to
OPEN_QUESTIONS.md - Never infer, never assume, never use default for critical fields
Authority Sources
- BRD §5: Personas (P1–P4)
- BRD §8: Phase A intake flow (Fast + Interview)
- BRD §10: Framework Packs (overlay compose + role mapping)
- BRD §11: BYO-LLM (optional enhancement)
Do NOT
- Use LLM for pack/role selection logic
- Infer answers to skipped questions
- Add questions not in the BRD schema
- Turn Fast mode into Interview mode via extra validation
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.

