allow1 - SKILL.md Agent Skill

name: allow1 description: "Use when generating, vectorizing, or converting images to SVG — logos, icons, illustrations, or 'allow1'/'quiver' references." allowed-tools: ["Read", "Write", "Edit", "Bash", "Glob", "Grep", "AskUserQuestion", "WebFetch"] argument-hint: "[generate|vectorize|batch] [prompt or image path]" user-invocable: false

ALLOW 1 (Quiver AI) — SVG Generation & Vectorization Skill

Uses the Quiver AI API to perform professional-level SVG generation from text prompts (Text-to-SVG) and raster image to SVG conversion (Image-to-SVG).

Prerequisites

ALLOW1_API_KEY environment variable is set
Base URL: https://api.quiver.ai/v1
Model: arrow-preview (current public model)

Quick Reference

Feature	Subcommand	Reference
Text-to-SVG Generation	`generate`	references/text-to-svg.md
Image-to-SVG Conversion	`vectorize`	references/image-to-svg.md
Batch Generation	`batch`	references/batch-workflow.md
API Reference	—	references/api-reference.md
Prompt Guide	—	references/prompt-guide.md

"Create SVG" / "Generate logo" / "Create icon" → Text-to-SVG
"Convert image to SVG" / "Vectorize" / "Trace" → Image-to-SVG
"allow1" / "quiver" → This skill in general
"Batch SVG" / "Multiple SVG" → Batch generation

Execution Flow

/allow1 [generate|vectorize|batch] [prompt or image]
    |
    +--[Step 1] Environment check
    |   +-- Verify $ALLOW1_API_KEY exists
    |   +-- Verify connection by fetching model list (GET /v1/models)
    |
    +--[Step 2] Requirements gathering (AskUserQuestion as needed)
    |   +-- Purpose: Logo / Icon / Illustration / UI parts
    |   +-- Style: Minimal / Flat / Line art / Gradient
    |   +-- Output count (n): 1-16 (default 4)
    |   +-- Save path
    |
    +--[Step 3] API execution
    |   +-- generate: POST /v1/svgs/generations
    |   +-- vectorize: POST /v1/svgs/vectorizations
    |   +-- batch: Execute multiple requests sequentially (rate limit aware)
    |
    +--[Step 4] Result processing
    |   +-- Save as SVG file
    |   +-- Token usage report
    |   +-- Retry as needed (429 / 5xx errors)
    |
    +--[Step 5] Optimization (optional)
        +-- Optimization suggestions via SVGO
        +-- Advice on viewBox / color adjustments

Basic Usage Examples

Text-to-SVG

# Simple logo generation
/allow1 generate "A minimalist mountain logo for a hiking app"

# With style instructions
/allow1 generate "Dashboard icon set" --style flat --n 8

Image-to-SVG

# Vectorize image file
/allow1 vectorize ./assets/logo.png

# Convert directly from URL
/allow1 vectorize https://example.com/image.png --auto-crop

Batch Generation

# Batch-generate multiple icons
/allow1 batch "home icon" "settings icon" "profile icon" "search icon"

API Call Patterns (curl)

Environment Variable Check + Connection Verification

# Verify API key
if [ -z "$ALLOW1_API_KEY" ]; then
  echo "ERROR: ALLOW1_API_KEY is not set"
  exit 1
fi

# Connection test (fetch model list)
curl -s https://api.quiver.ai/v1/models \
  -H "Authorization: Bearer $ALLOW1_API_KEY" | jq .

Text-to-SVG (Non-streaming)

curl -s https://api.quiver.ai/v1/svgs/generations \
  -H "Authorization: Bearer $ALLOW1_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "arrow-preview",
    "prompt": "A clean minimalist logo for a tech startup called Nexus",
    "instructions": "Use geometric shapes, monochrome palette, flat design",
    "n": 4,
    "temperature": 0.7,
    "stream": false
  }' | jq -r '.data[0].svg' > output.svg

Image-to-SVG

curl -s https://api.quiver.ai/v1/svgs/vectorizations \
  -H "Authorization: Bearer $ALLOW1_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "arrow-preview",
    "image": { "url": "https://example.com/logo.png" },
    "auto_crop": true,
    "target_size": 1024,
    "n": 1,
    "stream": false
  }' | jq -r '.data[0].svg' > vectorized.svg

Base64 Image Input

BASE64=$(base64 -i ./input.png)
curl -s https://api.quiver.ai/v1/svgs/vectorizations \
  -H "Authorization: Bearer $ALLOW1_API_KEY" \
  -H "Content-Type: application/json" \
  -d "{
    \"model\": \"arrow-preview\",
    \"image\": { \"base64\": \"$BASE64\" },
    \"auto_crop\": true,
    \"stream\": false
  }" | jq -r '.data[0].svg' > vectorized.svg

Rate Limiting and Error Handling

Limit	Value
Request limit	20 req / 60 sec (per organization)
Response headers	`X-RateLimit-Limit`, `X-RateLimit-Remaining`, `X-RateLimit-Reset`
On 429	Respect `Retry-After` header, exponential backoff
Billing	1 request = 1 credit (regardless of n value)

Error Code Quick Reference

HTTP	Code	Action
400	`invalid_request`	Check request parameters
401	`invalid_api_key`	Check `$ALLOW1_API_KEY`
402	`insufficient_credits`	Credit purchase required
403	`account_frozen`	Check account status
429	`rate_limit_exceeded`	Wait `Retry-After` seconds
500/502/503	Server error	Retry with exponential backoff

Related Skills

generate-slide — Slide Image Generation via Nano Banana Pro
generate-video — Video generation with Remotion
ui — UI component generation (for embedding SVG icons)