@quantod/qq 1.0.4 → 1.0.5

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@quantod/qq",
-  "version": "1.0.4",
+  "version": "1.0.5",
   "description": "Persistent queue for multi-stage Claude agent pipelines",
   "license": "UNLICENSED",
   "engines": {

package/src/mcp.ts

@@ -212,7 +212,7 @@ export async function startMcp(fixedPort?: number): Promise<void> {
 Available resources:
 - guide          — Complete reference: mental model, all MCP tools with parameters/returns/errors, SDK usage examples. Read this before using any QQ tool.
 - pipeline_design — Step-by-step pipeline design guide: how to design stages, write PIPELINES.md, and run the pre-implementation checklist.
-- chrome         — Chrome browser automation patterns for QQ pipelines: tab reuse, script injection, progress monitoring.`,
+- chrome         — Chrome browser automation patterns for QQ pipelines: read before using Chrome with pipelines.`,
     inputSchema: { resource: z.enum(RESOURCES) },
   }, ({ resource }) => {
     try { return ok(loadResource(resource)); }

package/src/resources/chrome.md

@@ -6,7 +6,7 @@ Read the `guide` resource first — this document covers Chrome-specific pattern
 
 ## How to get data out of Chrome
 
-Chrome devtools and connectors usually struggle to get large data packages from web pages in Chrome. Solution: download + load_file. The browser's download API has no mixed content restrictions. Trigger a download from the HTTPS page, then call the `load_file` MCP tool to ingest the file.
+Chrome devtools and connectors usually struggle to get large data packages from web pages in Chrome. Solution: download + load_file. The browser's download API has no mixed content restrictions. Trigger a download from the HTTPS page, then call the `load_file` MCP tool to ingest the file. Don't route large data through the context.
 
 **In the injected script (HTTPS page):**
 ```javascript

package/src/resources/guide.md

@@ -511,3 +511,9 @@ Use `replace: true` only when you own the payload as a structured object (e.g.,
 Before creating a new pipeline, read the `pipeline_design` resource. It walks through the design process, the PIPELINES.md template, a filled example, and a pre-implementation checklist.
 
 Document every pipeline in `PIPELINES.md` before pushing a single item.
+
+---
+
+## Chrome Browser Pipelines
+
+For pipelines that involve Chrome browser automation (scraping, navigation, form submission), read the `chrome` resource.