From syntactics-skills
Entry point for both the proposal and BA lifecycle at Syntactics Inc. Accepts any client input (brief, RFP, meeting notes, or approved proposal) and produces a single structured intake document used from proposal writing through database design. Trigger when a BA or Sales team member says "analyze client requirements", "start the intake", "we have a new project", "we have a new client", "I have client notes", "we got a proposal", or shares any client brief, email, RFP, or proposal PDF. Works in two modes: Pre-Proposal (from client brief - feeds proposal-grill and proposal-writer) and Post-Approval (from approved proposal - feeds database-designer). Always run before proposal-grill, database-designer, sprint-planner, and final-design. If the client has no brief or direction at all, run sync-client-discovery first.
How this skill is triggered — by the user, by Claude, or both
Slash command
/syntactics-skills:sync-project-intakeThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Read the client input. Produce the intake document. Do not design, spec, or estimate. Produce clarity only.
Read the client input. Produce the intake document. Do not design, spec, or estimate. Produce clarity only.
Proposal workflow: project-intake -> proposal-grill -> proposal-writer -> quotation BA design workflow: project-intake -> database-designer -> sprint-planner -> final-design
Step 1 - Project name: Ask "What is the project name?" - kebab-case, lowercase (e.g., client-portal). All output file paths use this name.
Step 2 - Mode detection: Ask:
"Are you working from a client brief / RFP / meeting notes (pre-proposal), or from a client-approved proposal (post-approval)?"
{project-name}-discovery.md from sync-client-discovery - intake status: Draft - next skill: proposal-grilldocs/sales/{project-name}-proposal.md) - intake status: Approved - next skill: database-designerAccept input as: PDF upload, pasted text, email thread, meeting notes, RFP document, or a discovery brief from sync-client-discovery.
Read everything provided. No output yet. While reading, note internally:
When the same concept appears under different names (e.g., "users" in one section and "members" in another), flag the inconsistency internally. Surface it in Phase 3 - do not silently pick one.
If more than 15 modules are found, show the first 15 in the Phase 2 table and summarize the remainder in a note below (e.g., "8 additional modules identified - listed in intake doc").
See references/extraction-rules.md for how to extract modules from unstructured client input.
Post inline in chat. Not a file yet.
| # | Module | Description (1 line) | Source | Confidence |
|---|---|---|---|---|
| 1 | User Authentication | Login, logout, password reset | Page 2 | Clear |
| 2 | Dashboard | Summary view for logged-in users | Implied | Inferred |
| 3 | Reports | Export data - format unspecified | Meeting notes | Ambiguous |
Confidence: Clear = fully described - Inferred = implied, not named - Ambiguous = mentioned but scope unclear
End with: X modules found. Y clear, Z ambiguous, N inferred.
Ask only what the input does not answer. See references/question-bank.md for the pool.
If zero ambiguous/inferred modules and all roles are defined: state no questions are needed and skip to Phase 4.
Rules: no generic questions - one sentence each - group by category - use Q-IDs - aim for 5-10 questions max, never more than 15 - prioritize ambiguous modules and missing role definitions first.
Sharpen fuzzy terms before closing questions. If the input uses a vague or overloaded term, include a sharpening question that proposes a precise interpretation. Do not leave it for grilling to catch.
Examples:
If answers reveal new gaps: run a follow-up Phase 3b with only the new questions before proceeding.
## Clarifying Questions
**Modules & Scope**
**User Roles**
**Business Rules**
**Data & Integrations**
**Technical**
**Timeline**
**Ambiguous or Inferred Modules**
Q-ID: [one-line question per category as needed]
Wait for answers before Phase 4.
After Q&A is answered (or user says "skip - proceed"):
File: docs/ba/{project-name}-intake.md
Follow references/output-format.md for the exact structure. Set Status to Draft (Pre-Proposal mode) or Approved (Post-Approval mode).
Artifact version frontmatter: Write this YAML block at the very top of the file before any other content.
Check if a previous version exists at the output path:
artifact_version: 1.0.0artifact_version, then bump:
1.0.0 -> 1.1.0)1.0.0 -> 1.0.1)---
artifact_version: {version}
generated_by: [email protected]
generated_at: {YYYY-MM-DD}
---
State the file path, then say:
Pre-Proposal mode:
Intake document ready. Next: proposal-grill - pass docs/ba/{project-name}-intake.md to stress-test scope before writing the proposal.
Post-Approval mode:
Intake document ready. Next: database-designer - pass docs/ba/{project-name}-intake.md to begin schema design.
references/extraction-rules.md - How to extract modules from unstructured client inputreferences/question-bank.md - Question pool for Phase 3references/output-format.md - Intake doc structurenpx claudepluginhub pbdevrepo/syntactics-skillsCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.