Skip to content
Back to skills

State

ASecurity

Create representation of current world state for a domain. Use when modeling system state, building world models, capturing entity relationships, or establishing baseline snapshots.

  • 4 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added February 8, 2026
ai-agentssqldockergitdatabase

Works with

  • cli

Security analysis

A100/100

Pro scans all 6 files and shows the line behind each finding

Scanned February 12, 2026

npx -y skills add synaptiai/agent-capability-standard --skill state --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of State?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for State
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/synaptiai-state/badge)](https://www.skillsdirectory.com/skills/synaptiai-state)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: state
description: Create representation of current world state for a domain. Use when modeling system state, building world models, capturing entity relationships, or establishing baseline snapshots.
argument-hint: "[scope] [schema] [timestamp]"
disable-model-invocation: false
user-invocable: true
allowed-tools: Read, Grep
context: fork
agent: explore
layer: MODEL
---

## Intent

Create a structured representation of the current state of a domain, system, or entity. This is the foundation for world modeling, enabling tracking of entities, relationships, and properties over time.

**Success criteria:**
- State captured in structured, queryable format
- Entities and relationships clearly identified
- Uncertainty and confidence explicitly represented
- Evidence anchors for all state assertions

**Compatible schemas:**
- `schemas/output_schema.yaml`
- `reference/world_state_schema.yaml`

## Inputs

| Parameter | Required | Type | Description |
|-----------|----------|------|-------------|
| `scope` | Yes | string | What domain/system to model (e.g., "user authentication", "payment processing") |
| `schema` | No | object | State schema defining expected structure |
| `timestamp` | No | string | Point in time for state (default: now) |
| `depth` | No | string | Modeling depth: surface, detailed, comprehensive |

## Procedure

1) **Define scope boundaries**: Clarify what is included in the state model
   - Identify system or domain boundaries
   - Determine entity types to include
   - Establish relationship types to capture

2) **Identify entities**: Enumerate entities within scope
   - Extract entity references from sources
   - Assign unique identifiers
   - Capture entity properties/attributes

3) **Map relationships**: Document connections between entities
   - Identify relationship types (contains, depends_on, references)
   - Note relationship directionality
   - Capture relationship properties

4) **Capture current values**: Record current state of each entity
   - Document property values with evidence
   - Note uncertainty where values are unclear
   - Record observation timestamp

5) **Apply schema**: Structure state according to schema if provided
   - Validate against schema constraints
   - Note schema violations or extensions needed
   - Ensure required fields are populated

6) **Ground state claims**: Attach evidence to all assertions
   - Reference source files, commands, or observations
   - Note confidence level for each assertion

## Output Contract

Return a structured object:

```yaml
state:
  world_id: string  # Unique identifier for this state snapshot
  scope: string  # Domain being modeled
  timestamp: string  # When state was captured
  entities:
    - id: string  # Unique entity identifier
      type: string  # Entity classification
      properties: object  # Entity attributes
      confidence: number  # Confidence in entity state
  relationships:
    - from: string  # Source entity ID
      to: string  # Target entity ID
      type: string  # Relationship type
      properties: object  # Relationship attributes
  meta:
    version: string  # State schema version
    completeness: number  # 0-1 estimate of coverage
confidence: 0..1  # Overall state model confidence
evidence_anchors: ["file:line", "observation:id"]
assumptions: []
```

### Field Definitions

| Field | Type | Description |
|-------|------|-------------|
| `state.world_id` | string | Unique identifier for state snapshot |
| `state.scope` | string | Domain or system being modeled |
| `state.entities` | array | Identified entities with properties |
| `state.relationships` | array | Connections between entities |
| `state.meta.completeness` | number | Estimated model completeness |
| `confidence` | number | Overall confidence in state accuracy |

## Examples

### Example 1: Model Authentication System State

**Input:**
```yaml
scope: "user authentication system"
depth: "detailed"
```

**Output:**
```yaml
state:
  world_id: "auth-state-20240115-001"
  scope: "user authentication system"
  timestamp: "2024-01-15T10:30:00Z"
  entities:
    - id: "auth-service"
      type: "service"
      properties:
        status: "running"
        version: "2.3.1"
        port: 8080
      confidence: 0.95
    - id: "user-db"
      type: "database"
      properties:
        type: "PostgreSQL"
        connection_pool: 20
        active_connections: 12
      confidence: 0.90
    - id: "session-cache"
      type: "cache"
      properties:
        type: "Redis"
        memory_used: "256MB"
        keys: 15420
      confidence: 0.90
  relationships:
    - from: "auth-service"
      to: "user-db"
      type: "depends_on"
      properties:
        connection_type: "direct"
        required: true
    - from: "auth-service"
      to: "session-cache"
      type: "depends_on"
      properties:
        connection_type: "direct"
        required: true
  meta:
    version: "1.0"
    completeness: 0.85
confidence: 0.88
evidence_anchors:
  - "config/services.yml:15-45"
  - "command:docker ps"
  - "command:redis-cli info"
assumptions:
  - "Service discovery reflects actual running state"
  - "Configuration matches deployed state"
```

### Example 2: Model Code Module State

**Input:**
```yaml
scope: "payment processing module"
schema:
  required_entities: ["class", "method", "dependency"]
```

**Output:**
```yaml
state:
  world_id: "payments-code-20240115"
  scope: "payment processing module"
  timestamp: "2024-01-15T11:00:00Z"
  entities:
    - id: "PaymentProcessor"
      type: "class"
      properties:
        file: "src/services/payment_processor.rb"
        lines: 145
        methods: 8
        complexity: 24
      confidence: 0.95
    - id: "process_payment"
      type: "method"
      properties:
        class: "PaymentProcessor"
        visibility: "public"
        params: ["order", "payment_method"]
        complexity: 12
      confidence: 0.95
    - id: "stripe-gem"
      type: "dependency"
      properties:
        name: "stripe"
        version: "8.0.0"
        usage: ["PaymentProcessor"]
      confidence: 0.90
  relationships:
    - from: "PaymentProcessor"
      to: "stripe-gem"
      type: "depends_on"
      properties:
        import_type: "require"
    - from: "process_payment"
      to: "PaymentProcessor"
      type: "member_of"
      properties:
        visibility: "public"
  meta:
    version: "1.0"
    completeness: 0.75
confidence: 0.85
evidence_anchors:
  - "src/services/payment_processor.rb:1-145"
  - "Gemfile:42"
assumptions:
  - "Static analysis reflects runtime behavior"
  - "No dynamic method definitions"
```

## Verification

- [ ] State includes world_id and timestamp
- [ ] All entities have unique IDs and types
- [ ] Relationships reference valid entity IDs
- [ ] Confidence scores present for entities
- [ ] Evidence anchors support state assertions

**Verification tools:** Read (to verify file references)

## Safety Constraints

- `mutation`: false
- `requires_checkpoint`: false
- `requires_approval`: false
- `risk`: low

**Capability-specific rules:**
- Do not modify state while modeling it
- Note when state may be stale or dynamic
- Flag entities with low confidence
- Do not invent entities without evidence

## Composition Patterns

**Commonly follows:**
- `observe` - Observations feed into state modeling
- `retrieve` - Retrieved data informs state
- `integrate` - Merged data forms state

**Commonly precedes:**
- `transition` - State enables transition modeling
- `compare` - States can be compared (diff)
- `simulate` - State is starting point for simulation

**Anti-patterns:**
- Never use state for predictions (use `predict`)
- Avoid state for single-value measurement (use `measure`)

**Workflow references:**
- See `reference/workflow_catalog.yaml#world_model_build` for state in world modeling
- See `reference/workflow_catalog.yaml#digital_twin_sync_loop` for state in digital twins

Files in this skill

  • SKILL.md7.7 KB
  • reference/event_schema.yaml3.5 KB
  • reference/transform_mapping_rawlog_to_observation.yaml2.2 KB
  • reference/workflow_catalog.yaml55.6 KB
  • reference/world_state_schema.yaml10.6 KB
  • schemas/output_schema.yaml726 B

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.