| name | scrunch-cli |
| description | Use this skill for service operations only. DO NOT use this skill for CLI implementation lifecycle work such as creating, testing, updating, troubleshooting, validating, removing, or documenting the CLI tool itself; delegate those tasks to cli-tool-expert. MANDATORY: Execute scrunch operations using the `scrunch` CLI tool. Provides a CLI interface for Scrunch AI API operations including brand visibility, AI search analytics, competitors, personas, prompts, and agent traffic. Triggers: scrunch, scrunch cli, scrunch brands, scrunch competitors, scrunch personas, scrunch prompts, scrunch query, scrunch responses, scrunch page audits, scrunch agent traffic, AI search visibility, brand presence. |
Provide expert guidance for using the `scrunch` CLI tool to interact with the Scrunch AI API. Covers brand management, competitor tracking, persona configuration, prompt management, aggregated query metrics, AI response inspection, page audits, and agent traffic analysis.
<quick_start>
scrunch auth login
scrunch auth status
scrunch brands list --table
scrunch query metrics 123 --fields date_week,brand_presence_percentage --start-date 2026-01-01 --end-date 2026-03-30
</quick_start>
<command_reference>
Auth:
scrunch auth login — Store API key
scrunch auth status — Check auth status using the shared profile JSON shape (auth profiles[].authenticated and auth profiles[].credential_types.api_key)
scrunch auth test — Test API connection
Brands:
scrunch brands list [--table] [--limit N] [--filter F] [--properties P]
scrunch brands get BRAND_ID [--table]
scrunch brands create --name NAME --website URL [--description DESC]
scrunch brands update BRAND_ID [--name NAME] [--website URL]
scrunch brands delete BRAND_ID
Competitors (nested under brand):
scrunch competitors list BRAND_ID [--table] [--limit N] [--filter F]
scrunch competitors get BRAND_ID COMPETITOR_ID [--table]
scrunch competitors create BRAND_ID --name NAME [--websites URL1,URL2]
scrunch competitors update BRAND_ID COMPETITOR_ID --name NAME
scrunch competitors delete BRAND_ID COMPETITOR_ID
Personas (nested under brand):
scrunch personas list BRAND_ID [--table] [--limit N] [--filter F]
scrunch personas get BRAND_ID PERSONA_ID [--table]
scrunch personas create BRAND_ID --name NAME --description DESC
scrunch personas update BRAND_ID PERSONA_ID --name NAME --description DESC
scrunch personas delete BRAND_ID PERSONA_ID
Prompts (nested under brand):
scrunch prompts list BRAND_ID [--table] [--limit N] [--filter F]
scrunch prompts get BRAND_ID PROMPT_ID [--table]
scrunch prompts create BRAND_ID --text TEXT --stage STAGE [--platforms P1,P2]
scrunch prompts delete BRAND_ID PROMPT_ID
Stages: Advice, Awareness, Evaluation, Comparison, Other
Platforms: chatgpt, claude, google_ai_overviews, perplexity, meta, google_ai_mode, google_gemini, copilot, grok
Query (aggregated metrics):
scrunch query metrics BRAND_ID --fields F1,F2 [--start-date DATE] [--end-date DATE] [--limit N] [--offset N] [--table]
Dimensions: date, date_week, date_month, date_quarter, date_year, prompt_id, prompt, persona_id, persona_name, ai_platform, competitor_id, competitor_name, branded, stage, prompt_topic, country
Metrics: responses, brand_presence_percentage, brand_position_score, brand_sentiment_score, competitor_presence_percentage, competitor_position_score, competitor_sentiment_score
Responses (AI response data):
scrunch responses list BRAND_ID [--platform P] [--prompt-id ID] [--persona-id ID] [--stage S] [--start-date D] [--end-date D] [--limit N] [--table]
Page Audits:
scrunch page-audits list BRAND_ID [--status S] [--url U] [--table]
scrunch page-audits get BRAND_ID AUDIT_ID [--table]
scrunch page-audits create BRAND_ID --urls URL1,URL2
Agent Traffic:
scrunch agent-traffic get BRAND_ID SITE_ID --start-date DATE --end-date DATE [--fields F] [--time-bucket B] [--table]
</command_reference>
<api_details>
Base URL: https://api.scrunchai.com/v1
Auth: Bearer token via SCRUNCH_API_KEY env var
Scopes: query, configure, create-brand
Pagination: limit + offset on list endpoints
Rate limiting: Honors Retry-After header with exponential backoff
</api_details>
<success_criteria>
- Command executes without errors
- Output is valid JSON (default) or formatted table (--table)
- Filters and limits are applied server-side where supported
</success_criteria>