io.github.flowpackplugin/flowpack
Measured conclusion: Deep-tested. L2 outcome: tools_ok.
Flowpack: Shopify design catalog — badge/button/bar looks, honest-urgency presets, diagnosis.
| Author | io.github.flowpackplugin |
|---|---|
| Access | remote |
| Publisher key required | No |
| Measured state | Deep-tested |
| L1 liveness | alive |
| L1 latency | 89 ms |
| Last measured | 2026-10-26 |
Add to your agent
Copy this into your MCP client config:
{
"mcpServers": {
"flowpack": {
"url": "https://flowpack-mcp.flowpackplugin.workers.dev/mcp",
"transport": "streamable-http"
}
}
}
Measured tools (5)
browseUniversal catalog reader. Routes by domain ('kits', 'presets', 'palettes', 'seasonal', 'settings', 'decorations', 'effects', 'diagnostics', 'anatomy', 'compatibility') with optional facet filter (e.g., 'button.hover') and free-text query. Returns paged items with id/name/description and metadata-only marker. Use this for ANY merchant question about 'what options are available' or 'show me X'.
Parameters:
domain(required) — stringpage— integerfacet— stringquery— stringpage_size— integer
Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared
planMASTER KEY tool. Analyzes natural-language merchant intent ('make my buttons feel premium', 'wellness brand vibe', 'BFCM aggressive') and returns a structured plan of settings to change across Flowpack's Button + Cart + Badge + Variant + Frosty Pill kits. Handles single tweaks (one hover style) and full overhauls (50+ settings) with the same call. Returns opaque plan_id, change count, preview URL, and reasoning. ALWAYS call preview() before execute().
Parameters:
intent(required) — string — Natural-language merchant intentcontext— object
Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared
explainNatural-language explanation of a plan_id (what it does) or setting_path (what one setting controls). Audience: merchant (default), developer, or reviewer. Detail: brief (1-2 sentences) or deep (full context + related settings + a11y notes). Use when merchant asks 'what does this do' or 'why this preset'.
Parameters:
detail— stringplan_id— stringaudience— stringsetting_path— string
Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared
recommendGoal-oriented recommendation engine. Scenarios: 'personalized_4_plans' (returns 4 diverse plans — safe / aspirational / seasonal / conversion-tested), 'preset_for_brand', 'color_palette_for_theme', 'touch_target_for_devices', 'motion_density_for_audience', 'seasonal_for_calendar', 'diagnostics_first_check'. Use for proactive suggestions when merchant intent is unclear.
Parameters:
scenario(required) — stringcontext— object
Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared
diagnoseRuntime + theme + settings harmony check. Free-text symptom (e.g., 'cart drawer feels slow', 'badges not appearing') OR kit-scoped check. Returns categorized issues (theme_compat / settings_conflict / runtime_error / a11y / performance) with severity and optional fix_plan_id callable via execute(). health_score 0-100.
Parameters:
kit— stringdeep— boolean — Run full theme compat sweep (slower)symptom— string
Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared
Measured by MCPYP probing. Untested fields are omitted, never estimated. Listing is always free; MCPYP never proxies calls or stores user tokens.