@oriro/orirocli 0.1.9 → 0.1.11

package/skills/technical/ai-product-builder/SKILL.md

@@ -1,180 +1,180 @@
----
-watermark: ORIRO
-name: ai-product-builder
-provider: ORIRO.ai
-copyright: Copyright (c) 2026 ORIRO.ai
-description: >
-
-
-
-  AI product development — LLM integration, RAG, prompt engineering, AI UX patterns, streaming, and building with AI APIs.
-
-
-  Sources: Stripe docs, Anthropic docs, Sanity/Contentful docs, industry-specific technical resources.
----
-
-# AI Product Development
-
-## LLM integration fundamentals
-
-### Calling the API (Anthropic)
-
-```ts
-import Anthropic from "@anthropic-ai/sdk";
-
-const client = new Anthropic({ apiKey: process.env.ANTHROPIC_API_KEY });
-
-const response = await client.messages.create({
-  model: "claude-opus-4-5", // or claude-sonnet-4-5, claude-haiku-4-5-20251001
-  max_tokens: 1024,
-  system: "You are a helpful assistant for a financial planning app.",
-  messages: [{ role: "user", content: userMessage }],
-});
-
-const text = response.content[0].type === "text" ? response.content[0].text : "";
-```
-
-### Streaming responses (essential for good UX)
-
-```ts
-// Long AI responses should stream — users see output as it generates
-const stream = await client.messages.create({
-  model: 'claude-sonnet-4-5',
-  max_tokens: 2048,
-  stream: true,
-  messages: [{ role: 'user', content: prompt }],
-});
-
-// In Next.js App Router:
-export async function POST(req: Request) {
-  const body = await req.json();
-  const encoder = new TextEncoder();
-  const stream = new ReadableStream({
-    async start(controller) {
-      const anthropicStream = client.messages.stream({ ... });
-      for await (const event of anthropicStream) {
-        if (event.type === 'content_block_delta' && event.delta.type === 'text_delta') {
-          controller.enqueue(encoder.encode(event.delta.text));
-        }
-      }
-      controller.close();
-    }
-  });
-  return new Response(stream, { headers: { 'Content-Type': 'text/plain; charset=utf-8' } });
-}
-```
-
-## RAG (Retrieval Augmented Generation)
-
-### Why RAG
-
-LLMs have knowledge cutoffs. RAG = search your own data, inject into context.
-Better than fine-tuning for: changing data, specific documents, proprietary knowledge.
-
-### RAG pipeline
-
-```
-User query → Embed query → Vector search → Top-k chunks → Augmented prompt → LLM → Answer
-```
-
-### Implementation
-
-```ts
-import OpenAI from "openai"; // or use Anthropic for embeddings
-import { Pinecone } from "@pinecone-database/pinecone";
-
-// 1. Embed user query
-const embedding = await openai.embeddings.create({
-  model: "text-embedding-3-small",
-  input: userQuery,
-});
-
-// 2. Search vector store
-const results = await pinecone.index("knowledge-base").query({
-  vector: embedding.data[0].embedding,
-  topK: 5,
-  includeMetadata: true,
-});
-
-// 3. Build augmented prompt
-const context = results.matches.map((m) => m.metadata.text).join("\n\n");
-const augmentedPrompt = `
-Context information:
-${context}
-
-User question: ${userQuery}
-
-Answer based on the context above. If the information isn't in the context, say so.
-`;
-
-// 4. Generate with context
-const response = await anthropic.messages.create({
-  model: "claude-sonnet-4-5",
-  max_tokens: 1024,
-  messages: [{ role: "user", content: augmentedPrompt }],
-});
-```
-
-## Tool use (function calling)
-
-```ts
-const tools = [
-  {
-    name: "get_weather",
-    description: "Get current weather for a location",
-    input_schema: {
-      type: "object",
-      properties: {
-        location: { type: "string", description: "City name or coordinates" },
-        units: { type: "string", enum: ["celsius", "fahrenheit"] },
-      },
-      required: ["location"],
-    },
-  },
-];
-
-const response = await client.messages.create({
-  model: "claude-sonnet-4-5",
-  max_tokens: 1024,
-  tools,
-  messages: [{ role: "user", content: "What's the weather in Paris?" }],
-});
-
-// Check if model wants to use a tool
-if (response.stop_reason === "tool_use") {
-  const toolUse = response.content.find((c) => c.type === "tool_use");
-  const result = await executeToolCall(toolUse.name, toolUse.input);
-  // Continue conversation with tool result
-}
-```
-
-## AI UX patterns
-
-**Streaming text:** Always stream for long responses. Users disengage if they wait.
-**Optimistic UI:** Show user message immediately, wait for AI response.
-**Stop generation:** Provide ability to stop mid-stream.
-**Retry / regenerate:** Always offer re-generation.
-**Citation / sources:** For RAG responses, cite the source documents.
-**Feedback:** Thumbs up/down on responses. Improves model or prompt over time.
-**Guardrails UI:** If request refused, explain why and offer alternatives.
-**Conversation history:** Maintain context across turns. Summarize long conversations.
-**Token awareness:** Show remaining context. Warn when approaching limits.
-
-## Prompt engineering
-
-**Be specific:** Vague prompts → vague outputs.
-**System prompt:** Set role, behavior, constraints, output format.
-**Few-shot examples:** 2-5 examples of desired input→output.
-**Chain of thought:** "Think step by step" or "Let's reason through this."
-**Output format:** "Respond in JSON", "Use markdown headers", "Maximum 3 bullet points."
-**Role assignment:** "You are an expert [role] who specializes in [domain]."
-
-## Cost management
-
-Claude Haiku: cheapest for simple tasks. Sonnet: balanced. Opus: most capable.
-Cache common prompts (Anthropic prompt caching).
-Set max_tokens appropriately. Don't over-generate.
-Log and monitor token usage per feature.
-Rate limit users to prevent runaway costs.
-
-Sources: Anthropic documentation (docs.anthropic.com — free), Vercel AI SDK (sdk.vercel.ai/docs — free), LangChain documentation (free), Building AI Products by Lenny Rachitsky (Substack)
+---
+watermark: ORIRO
+name: ai-product-builder
+provider: ORIRO.ai
+copyright: Copyright (c) 2026 ORIRO.ai
+description: >
+
+
+
+  AI product development — LLM integration, RAG, prompt engineering, AI UX patterns, streaming, and building with AI APIs.
+
+
+  Sources: Stripe docs, Anthropic docs, Sanity/Contentful docs, industry-specific technical resources.
+---
+
+# AI Product Development
+
+## LLM integration fundamentals
+
+### Calling the API (Anthropic)
+
+```ts
+import Anthropic from "@anthropic-ai/sdk";
+
+const client = new Anthropic({ apiKey: process.env.ANTHROPIC_API_KEY });
+
+const response = await client.messages.create({
+  model: "claude-opus-4-5", // or claude-sonnet-4-5, claude-haiku-4-5-20251001
+  max_tokens: 1024,
+  system: "You are a helpful assistant for a financial planning app.",
+  messages: [{ role: "user", content: userMessage }],
+});
+
+const text = response.content[0].type === "text" ? response.content[0].text : "";
+```
+
+### Streaming responses (essential for good UX)
+
+```ts
+// Long AI responses should stream — users see output as it generates
+const stream = await client.messages.create({
+  model: 'claude-sonnet-4-5',
+  max_tokens: 2048,
+  stream: true,
+  messages: [{ role: 'user', content: prompt }],
+});
+
+// In Next.js App Router:
+export async function POST(req: Request) {
+  const body = await req.json();
+  const encoder = new TextEncoder();
+  const stream = new ReadableStream({
+    async start(controller) {
+      const anthropicStream = client.messages.stream({ ... });
+      for await (const event of anthropicStream) {
+        if (event.type === 'content_block_delta' && event.delta.type === 'text_delta') {
+          controller.enqueue(encoder.encode(event.delta.text));
+        }
+      }
+      controller.close();
+    }
+  });
+  return new Response(stream, { headers: { 'Content-Type': 'text/plain; charset=utf-8' } });
+}
+```
+
+## RAG (Retrieval Augmented Generation)
+
+### Why RAG
+
+LLMs have knowledge cutoffs. RAG = search your own data, inject into context.
+Better than fine-tuning for: changing data, specific documents, proprietary knowledge.
+
+### RAG pipeline
+
+```
+User query → Embed query → Vector search → Top-k chunks → Augmented prompt → LLM → Answer
+```
+
+### Implementation
+
+```ts
+import OpenAI from "openai"; // or use Anthropic for embeddings
+import { Pinecone } from "@pinecone-database/pinecone";
+
+// 1. Embed user query
+const embedding = await openai.embeddings.create({
+  model: "text-embedding-3-small",
+  input: userQuery,
+});
+
+// 2. Search vector store
+const results = await pinecone.index("knowledge-base").query({
+  vector: embedding.data[0].embedding,
+  topK: 5,
+  includeMetadata: true,
+});
+
+// 3. Build augmented prompt
+const context = results.matches.map((m) => m.metadata.text).join("\n\n");
+const augmentedPrompt = `
+Context information:
+${context}
+
+User question: ${userQuery}
+
+Answer based on the context above. If the information isn't in the context, say so.
+`;
+
+// 4. Generate with context
+const response = await anthropic.messages.create({
+  model: "claude-sonnet-4-5",
+  max_tokens: 1024,
+  messages: [{ role: "user", content: augmentedPrompt }],
+});
+```
+
+## Tool use (function calling)
+
+```ts
+const tools = [
+  {
+    name: "get_weather",
+    description: "Get current weather for a location",
+    input_schema: {
+      type: "object",
+      properties: {
+        location: { type: "string", description: "City name or coordinates" },
+        units: { type: "string", enum: ["celsius", "fahrenheit"] },
+      },
+      required: ["location"],
+    },
+  },
+];
+
+const response = await client.messages.create({
+  model: "claude-sonnet-4-5",
+  max_tokens: 1024,
+  tools,
+  messages: [{ role: "user", content: "What's the weather in Paris?" }],
+});
+
+// Check if model wants to use a tool
+if (response.stop_reason === "tool_use") {
+  const toolUse = response.content.find((c) => c.type === "tool_use");
+  const result = await executeToolCall(toolUse.name, toolUse.input);
+  // Continue conversation with tool result
+}
+```
+
+## AI UX patterns
+
+**Streaming text:** Always stream for long responses. Users disengage if they wait.
+**Optimistic UI:** Show user message immediately, wait for AI response.
+**Stop generation:** Provide ability to stop mid-stream.
+**Retry / regenerate:** Always offer re-generation.
+**Citation / sources:** For RAG responses, cite the source documents.
+**Feedback:** Thumbs up/down on responses. Improves model or prompt over time.
+**Guardrails UI:** If request refused, explain why and offer alternatives.
+**Conversation history:** Maintain context across turns. Summarize long conversations.
+**Token awareness:** Show remaining context. Warn when approaching limits.
+
+## Prompt engineering
+
+**Be specific:** Vague prompts → vague outputs.
+**System prompt:** Set role, behavior, constraints, output format.
+**Few-shot examples:** 2-5 examples of desired input→output.
+**Chain of thought:** "Think step by step" or "Let's reason through this."
+**Output format:** "Respond in JSON", "Use markdown headers", "Maximum 3 bullet points."
+**Role assignment:** "You are an expert [role] who specializes in [domain]."
+
+## Cost management
+
+Claude Haiku: cheapest for simple tasks. Sonnet: balanced. Opus: most capable.
+Cache common prompts (Anthropic prompt caching).
+Set max_tokens appropriately. Don't over-generate.
+Log and monitor token usage per feature.
+Rate limit users to prevent runaway costs.
+
+Sources: Anthropic documentation (docs.anthropic.com — free), Vercel AI SDK (sdk.vercel.ai/docs — free), LangChain documentation (free), Building AI Products by Lenny Rachitsky (Substack)

package/skills/technical/analytics-setup/LICENSE

@@ -1,21 +1,21 @@
-MIT License
-
-Copyright (c) 2026 ORIRO.ai
-
-Permission is hereby granted, free of charge, to any person obtaining a copy
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all
-copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
+MIT License
+
+Copyright (c) 2026 ORIRO.ai
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/skills/technical/analytics-setup/SKILL.md

@@ -1,125 +1,125 @@
----
-watermark: ORIRO
-name: analytics-setup
-provider: ORIRO.ai
-copyright: Copyright (c) 2026 ORIRO.ai
-description: >
-
-  Analytics implementation — tracking setup, event design, funnel analysis, A/B testing, and data-driven product decisions.
-
-
-
-
-
-
-
-
-
-
-  Sources: Google Search Central, GDPR text, AWS security documentation, Stripe API docs.
----
-
-# Analytics Implementation
-
-## Event tracking design
-
-### Naming convention
-
-```
-Object Action: user_signed_up, post_published, payment_completed, button_clicked
-```
-
-Consistent naming from day one. Inconsistency = unusable data at scale.
-
-### What to track
-
-**Product events:**
-
-- Acquisition: page_viewed (first session), signup_started, signup_completed, onboarding_step_completed
-- Activation: first_key_action (varies by product — first post created, first file uploaded, first transaction)
-- Retention: session_started, core_feature_used (daily/weekly)
-- Revenue: plan_selected, trial_started, subscription_created, payment_failed
-- Referral: share_clicked, invite_sent, referral_link_clicked
-
-**Don't track everything.** Track actions that matter. Every event that's not used is noise.
-
-## Implementation (Segment / Posthog)
-
-### Segment
-
-```ts
-// Track event
-analytics.track("post_published", {
-  userId: user.id,
-  postId: post.id,
-  category: post.category,
-  wordCount: post.wordCount,
-  publishedAt: new Date().toISOString(),
-});
-
-// Identify user
-analytics.identify(user.id, {
-  email: user.email,
-  plan: user.subscription.plan,
-  createdAt: user.createdAt,
-  company: user.company?.name,
-});
-```
-
-### PostHog (self-hosted or cloud)
-
-```ts
-import posthog from "posthog-js";
-
-// Initialize
-posthog.init(process.env.NEXT_PUBLIC_POSTHOG_KEY, {
-  api_host: process.env.NEXT_PUBLIC_POSTHOG_HOST,
-  loaded: (posthog) => {
-    if (process.env.NODE_ENV === "development") posthog.opt_out_capturing();
-  },
-});
-
-// Track
-posthog.capture("post_published", { postId: post.id, category: post.category });
-
-// Identify
-posthog.identify(userId, { email, plan });
-```
-
-## Funnel analysis
-
-```
-Signup page view → Form submitted → Email verified → Onboarding started → Onboarding completed → Core action taken → Subscription
-Step 1: 10,000 → Step 2: 5,000 (50%) → Step 3: 3,500 (70%) → Step 4: 2,800 (80%) → Step 5: 1,400 (50%) → Step 6: 840 (60%) → Step 7: 168 (20%)
-```
-
-Where is the biggest drop? That's where to focus.
-
-## A/B testing
-
-```ts
-// PostHog feature flags for A/B tests
-const variant = posthog.getFeatureFlag("checkout-cta-test");
-// variant = 'control' or 'test'
-
-// Show different UI based on variant
-const ctaText = variant === "test" ? "Start free trial" : "Get started";
-```
-
-**Don't peek at results early** — statistical significance requires adequate sample size.
-Calculate sample size before starting: need ~95% confidence, 80% power.
-
-## Key metrics by product type
-
-**SaaS:** DAU/WAU/MAU, DAU/MAU ratio (engagement), Activation rate, NRR.
-**E-commerce:** Conversion rate, Cart abandonment, Average order value, Return rate.
-**Content:** Page views, Time on page, Pages per session, Scroll depth, Return visitors.
-**Marketplace:** Liquidity (listings → transactions), GMV, Take rate, NPS both sides.
-
-## Data quality
-
-**Source of truth:** Agree on ONE definition for each metric. "Active user" must mean the same thing everywhere.
-**Validate tracking:** Use Segment debugger, PostHog event explorer, or browser console to verify events fire correctly.
-**Data warehouse:** Segment → Snowflake/BigQuery/Redshift for complex analysis, not just dashboards.
-
-Sources: Segment documentation (segment.com/docs — free), PostHog documentation (posthog.com/docs — free), Google Analytics 4 documentation (free)
+---
+watermark: ORIRO
+name: analytics-setup
+provider: ORIRO.ai
+copyright: Copyright (c) 2026 ORIRO.ai
+description: >
+
+  Analytics implementation — tracking setup, event design, funnel analysis, A/B testing, and data-driven product decisions.
+
+
+
+
+
+
+
+
+
+
+  Sources: Google Search Central, GDPR text, AWS security documentation, Stripe API docs.
+---
+
+# Analytics Implementation
+
+## Event tracking design
+
+### Naming convention
+
+```
+Object Action: user_signed_up, post_published, payment_completed, button_clicked
+```
+
+Consistent naming from day one. Inconsistency = unusable data at scale.
+
+### What to track
+
+**Product events:**
+
+- Acquisition: page_viewed (first session), signup_started, signup_completed, onboarding_step_completed
+- Activation: first_key_action (varies by product — first post created, first file uploaded, first transaction)
+- Retention: session_started, core_feature_used (daily/weekly)
+- Revenue: plan_selected, trial_started, subscription_created, payment_failed
+- Referral: share_clicked, invite_sent, referral_link_clicked
+
+**Don't track everything.** Track actions that matter. Every event that's not used is noise.
+
+## Implementation (Segment / Posthog)
+
+### Segment
+
+```ts
+// Track event
+analytics.track("post_published", {
+  userId: user.id,
+  postId: post.id,
+  category: post.category,
+  wordCount: post.wordCount,
+  publishedAt: new Date().toISOString(),
+});
+
+// Identify user
+analytics.identify(user.id, {
+  email: user.email,
+  plan: user.subscription.plan,
+  createdAt: user.createdAt,
+  company: user.company?.name,
+});
+```
+
+### PostHog (self-hosted or cloud)
+
+```ts
+import posthog from "posthog-js";
+
+// Initialize
+posthog.init(process.env.NEXT_PUBLIC_POSTHOG_KEY, {
+  api_host: process.env.NEXT_PUBLIC_POSTHOG_HOST,
+  loaded: (posthog) => {
+    if (process.env.NODE_ENV === "development") posthog.opt_out_capturing();
+  },
+});
+
+// Track
+posthog.capture("post_published", { postId: post.id, category: post.category });
+
+// Identify
+posthog.identify(userId, { email, plan });
+```
+
+## Funnel analysis
+
+```
+Signup page view → Form submitted → Email verified → Onboarding started → Onboarding completed → Core action taken → Subscription
+Step 1: 10,000 → Step 2: 5,000 (50%) → Step 3: 3,500 (70%) → Step 4: 2,800 (80%) → Step 5: 1,400 (50%) → Step 6: 840 (60%) → Step 7: 168 (20%)
+```
+
+Where is the biggest drop? That's where to focus.
+
+## A/B testing
+
+```ts
+// PostHog feature flags for A/B tests
+const variant = posthog.getFeatureFlag("checkout-cta-test");
+// variant = 'control' or 'test'
+
+// Show different UI based on variant
+const ctaText = variant === "test" ? "Start free trial" : "Get started";
+```
+
+**Don't peek at results early** — statistical significance requires adequate sample size.
+Calculate sample size before starting: need ~95% confidence, 80% power.
+
+## Key metrics by product type
+
+**SaaS:** DAU/WAU/MAU, DAU/MAU ratio (engagement), Activation rate, NRR.
+**E-commerce:** Conversion rate, Cart abandonment, Average order value, Return rate.
+**Content:** Page views, Time on page, Pages per session, Scroll depth, Return visitors.
+**Marketplace:** Liquidity (listings → transactions), GMV, Take rate, NPS both sides.
+
+## Data quality
+
+**Source of truth:** Agree on ONE definition for each metric. "Active user" must mean the same thing everywhere.
+**Validate tracking:** Use Segment debugger, PostHog event explorer, or browser console to verify events fire correctly.
+**Data warehouse:** Segment → Snowflake/BigQuery/Redshift for complex analysis, not just dashboards.
+
+Sources: Segment documentation (segment.com/docs — free), PostHog documentation (posthog.com/docs — free), Google Analytics 4 documentation (free)