| name | init |
| description | Initialize structured context for this project. Scaffolds .scs/ with all 11 concern bundles, scans the project, and recommends where to start. |
| argument-hint | [project-name] |
| allowed-tools | Read, Glob, Grep, Write, Bash(ls *), Bash(mkdir -p *), Bash(cat package.json), Bash(cat requirements.txt), Bash(cat go.mod), Bash(cat Cargo.toml), Bash(cat pyproject.toml) |
Invocation Rules
- User-initiated only: Do NOT invoke this skill unless the user explicitly runs
/scs-team:init. Never auto-invoke based on project state or conversation context.
- Confirm before writing: After scanning the project, present a summary of the
.scs/ scaffold you plan to create and get explicit user confirmation before writing any files.
SCS Team Init - Initialize Structured Context
You are scaffolding a new .scs/ directory for this project. This is the entry point for teams adopting structured context.
Your Process
Step 1: Detect Project Info
Scan the project to understand what exists:
Language/Framework Detection (check in order):
package.json → Node.js (check for Next.js, React, Express, etc.)
requirements.txt or pyproject.toml or setup.py → Python (check for FastAPI, Django, Flask)
go.mod → Go
Cargo.toml → Rust
pom.xml or build.gradle → Java/Kotlin
Gemfile → Ruby
.csproj or .sln → .NET
Existing Documentation (scan for):
docs/, doc/, documentation/ directories
- Root-level:
PRD.md, ARCHITECTURE.md, SECURITY.md, README.md
- Any
*.md files that look like project documentation
CLAUDE.md, .claude/rules/ (existing Claude context)
Compliance Signals (look for keywords in docs and code):
- HIPAA: "PHI", "protected health information", "HIPAA", "healthcare"
- SOC2: "SOC2", "SOC 2", "trust service criteria"
- PCI: "PCI", "cardholder", "payment card"
- GDPR: "GDPR", "personal data", "data subject"
- CHAI: "clinical AI", "health AI", "CHAI"
Step 2: Determine Project Name
Use the argument if provided. Otherwise:
- Check
package.json name field
- Check
pyproject.toml project name
- Check
go.mod module name
- Fall back to directory name
Step 3: Create .scs/ Scaffold
Create the following structure:
.scs/
├── project.yaml # Project bundle manifest
├── concerns/
│ ├── architecture.yaml
│ ├── security.yaml
│ ├── performance.yaml
│ ├── usability.yaml
│ ├── compliance.yaml
│ ├── data.yaml
│ ├── testing.yaml
│ ├── deployment.yaml
│ ├── safety.yaml
│ ├── ethics.yaml
│ └── business.yaml
└── scds/ # Empty - populated by add/draft/use
Project Bundle (project.yaml):
id: bundle:<project-name>
type: project
version: "DRAFT"
title: "<Project Name>"
description: "<Brief description from README or detected context>"
imports:
- bundle:architecture
- bundle:security
- bundle:performance
- bundle:usability
- bundle:compliance
- bundle:data
- bundle:testing
- bundle:deployment
- bundle:safety
- bundle:ethics
- bundle:business
scds: []
provenance:
created_by: "scs-team init"
created_at: "<ISO timestamp>"
rationale: "Project bundle scaffolded by scs-team init"
Concern Bundles (each concern follows this pattern):
id: bundle:<concern-slug>
type: concern
version: "DRAFT"
title: "<Project Name> - <Concern Title>"
description: "<Brief description of what this concern covers>"
imports: []
scds: []
provenance:
created_by: "scs-team init"
created_at: "<ISO timestamp>"
rationale: "Scaffolded by scs-team init"
Use these exact concern definitions:
| # | Slug | Title | Description | Expected SCDs |
|---|
| 1 | architecture | Architecture | System structure, components, technology stack, and integration points | system-context, tech-stack, integration-map, component-model |
| 2 | security | Security | Authentication, authorization, data protection, and threat modeling | authn-authz, data-protection, data-handling, threat-model |
| 3 | performance | Performance & Reliability | Response times, availability targets, fault tolerance, and scaling strategy | response-time, availability, fault-tolerance, scalability |
| 4 | usability | Usability & Accessibility | UX principles, accessibility compliance, and error handling | ux-principles, accessibility-compliance, error-handling-ux |
| 5 | compliance | Compliance & Governance | Regulatory requirements, audit controls, and policy enforcement | hipaa-compliance, soc2-controls, audit-requirements, policy-enforcement |
| 6 | data | Data & Provenance | Data models, lineage tracking, and retention policies | data-model, provenance-tracking, retention-policy |
| 7 | testing | Testing & Validation | Test coverage, validation plans, and QA procedures | test-coverage, validation-plan, qa-procedures |
| 8 | deployment | Deployment & Operations | Infrastructure, observability, and incident response | infrastructure-definition, observability, incident-response |
| 9 | safety | Safety & Risk | Risk assessments and safety checklists | risk-assessment, safety-checklist |
| 10 | ethics | Ethics & AI Accountability | AI usage policies, audit trails, and bias detection | ai-usage-policy, audit-trail, model-bias |
| 11 | business | Business Context | Problem definition, stakeholders, objectives, opportunity analysis, constraints, and success criteria | problem-definition, stakeholders, business-objectives, opportunity-analysis, constraints-and-assumptions, success-criteria |
Step 4: Report Findings and Recommend Next Steps
Present a summary:
## SCS Initialized
**Project**: <name>
**Language**: <detected language/framework>
**Location**: .scs/ (11 concern bundles scaffolded)
### What I Found
<table of detected docs, compliance signals, existing context>
### Recommended Next Steps
<prioritized list based on what was found>
Recommendation Logic:
-
Found existing documentation → Start with /scs-team:add:
/scs-team:add ./docs/PRD.md
/scs-team:add ./docs/architecture.md
-
No documentation found → Start with drafting:
/scs-team:draft business # Start with the "why"
/scs-team:draft architecture # Then the "how"
-
Compliance keywords detected → Add standards:
/scs-team:use hipaa # Healthcare detected
/scs-team:use soc2 # Security controls detected
-
Existing CLAUDE.md found → Note it will be enhanced (not replaced):
Existing CLAUDE.md found - scs-team will add a managed section
(your existing content will not be modified)
Important Rules
- Do NOT generate any SCDs - init only creates the scaffold
- Do NOT modify existing files - only create new ones in
.scs/
- Do NOT overwrite if
.scs/ already exists - warn the user and show what's there
- Keep it fast - scan, scaffold, report. Don't ask questions during init.
- The
scds: [] arrays in concern bundles are intentionally empty - they get populated by add, draft, and use
If .scs/ Already Exists
## SCS Already Initialized
Found existing .scs/ directory with:
- <N> concern bundles
- <N> SCDs
- Project bundle: <version status>
Run `/scs-team:status` to see the full picture.
To start fresh, delete .scs/ and run `/scs-team:init` again.
Example Interaction
User: /scs-team:init
You: "Let me scan your project..."
[Scan project files, docs, dependencies]
You: "## SCS Initialized
Project: patient-portal
Language: Python (FastAPI)
Location: .scs/ (11 concern bundles scaffolded)
What I Found
| Item | Details |
|---|
| Language | Python 3.11 (FastAPI, SQLAlchemy) |
| Docs | docs/PRD.md, docs/api-spec.md |
| Compliance | HIPAA keywords found in README |
| Existing Context | CLAUDE.md exists (will be enhanced, not replaced) |
Recommended Next Steps
-
Process your docs (you have existing documentation):
/scs-team:add docs/PRD.md
/scs-team:add docs/api-spec.md
-
Add HIPAA compliance (healthcare keywords detected):
/scs-team:use hipaa
-
Draft what's missing (no architecture doc found):
/scs-team:draft architecture
/scs-team:draft security
-
Check progress:
/scs-team:status
```"