chore(release): publish v0.7.0 (#26)

Co-authored-by: ct-sdks[bot] <153784748+ct-sdks[bot]@users.noreply.github.com>

Author: ct-sdks[bot]

.agents/plugins/commercetools/.codex-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "commercetools",
   "displayName": "commercetools",
-  "version": "0.6.0",
+  "version": "0.7.0",
   "description": "Build commercetools solutions faster",
   "author": {
     "name": "commercetools",

.agents/plugins/commercetools/skills/commercetools-checkout/SKILL.md

@@ -30,14 +30,14 @@ Three modes — choose based on how much of the checkout flow Checkout should ow
 
 When this skill is invoked, always follow these steps:
 
-1. **Gather context (required, run first)** — Always begin by gathering context for this skill. This is the mandatory grounding step: it gathers the latest verified documentation as context for you (the agent). **Do not skip it, and do not replace it with another tool** (such as an MCP documentation-search tool) — run this command:
+1. **Docs search (required, run first)** — Always begin by searching docs for this skill. This is the mandatory grounding step: it gathers the latest verified documentation as context for you (the agent). **Do not skip it, and do not replace it with another tool** (such as an MCP documentation-search tool) This script optimizes for tuned search results  — run this command:
    ```bash
-   node scripts/gather-context.mjs \
+   node scripts/docs-search.mjs \
      --query "<extract key terms from user's question>" \
-     --client-name "<current-client>" \
+     --app-name "<current-app ex: claude, copilot, codex>" \
      --model "<current-model>" \
      --skill-name "commercetools-checkout" \
-     --limit 3
+     --limit 20
    ```
    Use its output as your primary grounding. You *may additionally* use other tools (such as the commercetools documentation MCP) for deeper, follow-up search.
 

.agents/plugins/commercetools/skills/commercetools-checkout/scripts/docs-search.mjs

@@ -0,0 +1,151 @@
+#!/usr/bin/env node
+
+/**
+ * Search docs for the agent
+ *
+ * Gathers relevant commercetools documentation as grounding context for the agent, with instrumentation headers.
+ * Results are written to stdout for consumption by AI agents.
+ * 
+ * Usage:
+ *   node scripts/docs-search.mjs --query "how to create payment" --app-name <name> --model <model> --skill-name <skill> [options]
+ * 
+ * Required:
+ *   --query <string>          Search query
+ *   --app-name <string>       App identifier (e.g., vscode-copilot, claude-desktop)
+ *   --model <string>          Model name (e.g., claude-sonnet-4.5, gpt-4)
+ *   --skill-name <string>     Skill identifier (e.g., commercetools-checkout)
+ * 
+ * Optional:
+ *   --limit <number>          Number of results (default: 3)
+ *   --content-types <string>  Comma-separated content types
+ */
+
+import { parseArgs } from 'util';
+
+const CONTEXT_URL = 'https://docs.commercetools.com/apis/rest/tools/documentation-search';
+
+// Parse command line arguments
+const { values } = parseArgs({
+  options: {
+    query: { type: 'string' },
+    limit: { type: 'string', default: '3' },
+    'app-name': { type: 'string' },
+    model: { type: 'string' },
+    'skill-name': { type: 'string' },
+    'content-types': { type: 'string' },
+  },
+});
+
+// Validate required parameters
+const missingParams = [];
+if (!values.query) missingParams.push('--query');
+if (!values['app-name']) missingParams.push('--app-name');
+if (!values.model) missingParams.push('--model');
+if (!values['skill-name']) missingParams.push('--skill-name');
+
+if (missingParams.length > 0) {
+  console.error(`Error: Missing required parameters: ${missingParams.join(', ')}`);
+  console.error('Usage: node scripts/docs-search.mjs --query "search" --app-name "app" --model "model" --skill-name "skill"');
+  process.exit(1);
+}
+
+const normalizedLimit = Math.min(parseInt(values.limit, 10), 20)
+
+// Build request body
+const requestBody = {
+  query: values.query,
+  limit: normalizedLimit,
+  products: ['Composable Commerce', 'Checkout', 'Connect', 'AI Hub']
+};
+
+// Add optional content types filter
+if (values['content-types']) {
+  requestBody.contentTypes = values['content-types'].split(',').map(t => t.trim());
+}
+
+const requestData = JSON.stringify(requestBody);
+
+async function main() {
+  const controller = new AbortController();
+  const timeoutId = setTimeout(() => controller.abort(), 60000);
+
+  try {
+    const res = await fetch(CONTEXT_URL, {
+      method: 'POST',
+      headers: {
+        'Content-Type': 'application/json',
+        'Content-Length': Buffer.byteLength(requestData),
+        'User-Agent': `${values['app-name']}/1.0 (${values.model})`,
+        'X-Model': values.model,
+        'X-Client-Type': values['app-name'],
+        'X-Skill-Name': values['skill-name'],
+      },
+      body: requestData,
+      signal: controller.signal,
+    });
+
+    if (res.status !== 200) {
+      process.exit(0);
+    }
+
+    let response;
+    try {
+      response = await res.json();
+    } catch (err) {
+      process.exit(0);
+    }
+
+    if (response.error) {
+      process.exit(0);
+    }
+
+    // Format results for AI agent consumption
+    if (response.result && Array.isArray(response.result)) {
+      const results = response.result;
+
+      if (results.length === 0) {
+        process.stdout.write('No results found. Try broadening your search query or adjusting filters.\n');
+        process.exit(0);
+      }
+
+      // Write results in markdown format
+      process.stdout.write(`# Documentation Search Results\n\n`);
+      process.stdout.write(`Query: "${values.query}"\n`);
+      process.stdout.write(`Found ${results.length} result(s)\n\n`);
+      process.stdout.write('---\n\n');
+
+      results.forEach((item, index) => {
+        process.stdout.write(`## Result ${index + 1}\n\n`);
+
+        if (item.metadata) {
+          if (item.metadata.title) {
+            process.stdout.write(`**Title:** ${item.metadata.title}\n\n`);
+          }
+          if (item.metadata.url) {
+            const url = item.metadata.url.split('#')[0];
+            process.stdout.write(`**URL:** ${url}\n\n`);
+          }
+          if (item.metadata.contentType) {
+            process.stdout.write(`**Type:** ${item.metadata.contentType}\n\n`);
+          }
+        }
+
+        if (item.content) {
+          process.stdout.write(`**Content:**\n\n${item.content}\n\n`);
+        }
+
+        process.stdout.write('---\n\n');
+      });
+    } else {
+      // Fallback: output raw response
+      process.stdout.write(JSON.stringify(response, null, 2));
+      process.stdout.write('\n');
+    }
+  } catch (err) {
+    process.exit(0);
+  } finally {
+    clearTimeout(timeoutId);
+  }
+}
+
+main();

.agents/plugins/commercetools/skills/commercetools-checkout/scripts/gather-context.mjs

@@ -24,14 +24,14 @@ metadata:
 
 When this skill is invoked, always follow these steps:
 
-1. **Gather context (required, run first)** — Always begin by gathering context for this skill. This is the mandatory grounding step: it gathers the latest verified documentation as context for you (the agent). **Do not skip it, and do not replace it with another tool** (such as an MCP documentation-search tool) — run this command:
+1. **Docs search (required, run first)** — Always begin by searching docs for this skill. This is the mandatory grounding step: it gathers the latest verified documentation as context for you (the agent). **Do not skip it, and do not replace it with another tool** (such as an MCP documentation-search tool) This script optimizes for tuned search results  — run this command:
    ```bash
-   node scripts/gather-context.mjs \
+   node scripts/docs-search.mjs \
      --query "<extract key terms from user's question>" \
-     --client-name "<current-client>" \
+     --app-name "<current-app ex: claude, copilot, codex>" \
      --model "<current-model>" \
      --skill-name "commercetools-commerce-patterns" \
-     --limit 3
+     --limit 10
    ```
    Use its output as your primary grounding. You *may additionally* use other tools (such as the commercetools documentation MCP) for deeper, follow-up search.