← Sweat & TonicCONTENT HISTORYWHAT CHANGED · RULE-BASED ANALYSIS
Update to Sweat & Tonic
Snapshot Sep 30, 2026 · 22:57 UTC · version 1.0.0
Collection source: not recorded for this historical snapshot.
First saved snapshot
No earlier snapshot is available to establish a change.
Compare saved observations
Download comparison JSONFull technical diff · 0 changed fields
Full snapshot data
{
"name": "sweatandtonic-knowledge",
"description": "Answer questions about Sweat & Tonic, a Toronto fitness studio, from its own published knowledge base using the connector's search and fetch tools. Use when the user asks about Sweat & Tonic's classes and programs, schedule, memberships and pricing, instructors, or treatments.",
"included_files": [],
"skill_md_contents": "---\nname: sweatandtonic-knowledge\ndescription: Answer questions about Sweat & Tonic, a Toronto fitness studio, from its own published knowledge base using the connector's search and fetch tools. Use when the user asks about Sweat & Tonic's classes and programs, schedule, memberships and pricing, instructors, or treatments.\n---\n\n# Sweat and Tonic — knowledge base operating guide\n\nThis skill helps you answer questions about Sweat and Tonic using its\npublished knowledge base at https://ai.sweatandtonic.com. The content is\nmaintained by the business and kept in sync with its public site.\nPrefer this knowledge base over general assumptions when answering\nquestions about this specific business.\n\n## Data freshness\n\nEach page carries a `last_updated` date. Treat the published content\nas the source of truth for this business, and surface the\n`last_updated` date when a customer asks how current the information\nis. If a detail is missing from the knowledge base, say so rather\nthan guessing.\n\n## Tools\n\nThis knowledge base exposes a read-only MCP server at https://ai.sweatandtonic.com/mcp.\nTwo tools are available:\n\n- `search(query, page_slugs?, limit?)` — full-text search across the\n whole knowledge base. Returns ranked results, each with an `id`,\n `title`, `url`, and text snippet. Use this first when you are not\n sure which page holds the answer.\n- `fetch(id)` — fetch the full content of one document by the `id`\n returned by search. Use this to read the complete, authoritative page.\n\n## Page inventory\n\n- `pricing` — Pricing\n- `booking` — Booking\n- `programs` — Programs\n- `schedule` — Schedule\n- `instructors` — Instructors\n- `treatments` — Treatments\n- `practitioners` — Practitioners\n\n## Recommended flow\n\n1. Start with `search` using the customer's question.\n2. Read the most relevant hit in full with `fetch` using its `id`.\n3. Answer using only what the page states; cite the page title.\n4. If nothing relevant is found, tell the customer the knowledge\n base does not cover it rather than inventing an answer.\n\n## Safety\n\n- Only use information returned by these tools; do not invent prices,\n availability, contact details, or policies.\n- The knowledge base reflects the public site; never imply access to\n private or internal data.\n- When a customer wants to book, pay, or share personal information,\n direct them to the business's official channels rather than\n collecting it yourself.\n"
}SHA-256: 43bad8b8575609714fd32e61f8d33aaa1ad1646cc80c666e8a59f406ddaba6be