@apmantza/greedysearch-pi 1.4.1 → 1.4.2

package/launch.mjs

@@ -5,8 +5,8 @@
 // the "Allow remote debugging?" dialog entirely. It runs on port 9222 so it doesn't
 // conflict with your main Chrome session (which may use port 9223).
 //
-// search.mjs passes CDP_PROFILE_DIR so cdp.mjs targets this dedicated Chrome
-// without ever touching the user's main Chrome DevToolsActivePort file.
+// search.mjs passes CDP_PROFILE_DIR so cdp.mjs targets this dedicated Chrome
+// without ever touching the user's main Chrome DevToolsActivePort file.
 //
 // Usage:
 //   node launch.mjs          — launch (or report if already running)
@@ -14,8 +14,8 @@
 //   node launch.mjs --status — check if running
 
 import { spawn } from 'child_process';
-import { existsSync, writeFileSync, readFileSync, mkdirSync, unlinkSync } from 'fs';
-import { tmpdir, platform } from 'os';
+import { existsSync, writeFileSync, readFileSync, mkdirSync, unlinkSync } from 'fs';
+import { tmpdir, platform } from 'os';
 import { join } from 'path';
 import http from 'http';
 
@@ -42,8 +42,8 @@ function findChrome() {
   return candidates.find(existsSync) || null;
 }
 
-const CHROME_FLAGS = [
-  `--remote-debugging-port=${PORT}`,
+const CHROME_FLAGS = [
+  `--remote-debugging-port=${PORT}`,
   '--disable-features=DevToolsPrivacyUI',      // suppresses "Allow remote debugging?" dialog
   '--no-first-run',
   '--no-default-browser-check',
@@ -97,21 +97,21 @@ async function writePortFile(timeoutMs = 15000) {
   return false;
 }
 
-// ---------------------------------------------------------------------------
+// ---------------------------------------------------------------------------
 
 async function main() {
   const arg = process.argv[2];
 
-  if (arg === '--kill') {
-    const pid = isRunning();
-    if (pid) {
-      try { process.kill(pid, 'SIGTERM'); console.log(`Stopped Chrome (pid ${pid}).`); }
-      catch (e) { console.error(`Failed: ${e.message}`); }
-    } else {
-      console.log('GreedySearch Chrome is not running.');
-    }
-    return;
-  }
+  if (arg === '--kill') {
+    const pid = isRunning();
+    if (pid) {
+      try { process.kill(pid, 'SIGTERM'); console.log(`Stopped Chrome (pid ${pid}).`); }
+      catch (e) { console.error(`Failed: ${e.message}`); }
+    } else {
+      console.log('GreedySearch Chrome is not running.');
+    }
+    return;
+  }
 
   if (arg === '--status') {
     const pid = isRunning();
@@ -123,12 +123,12 @@ async function main() {
   // Already running?
   const existing = isRunning();
   if (existing) {
-    const ready = await writePortFile(5000);
-    if (ready) {
-      console.log(`GreedySearch Chrome already running (pid ${existing}, port ${PORT}).`);
-      console.log('Dedicated GreedySearch DevToolsActivePort is ready.');
-      return;
-    }
+    const ready = await writePortFile(5000);
+    if (ready) {
+      console.log(`GreedySearch Chrome already running (pid ${existing}, port ${PORT}).`);
+      console.log('Dedicated GreedySearch DevToolsActivePort is ready.');
+      return;
+    }
     // Stale PID — process alive but not Chrome on port 9223. Fall through to fresh launch.
     console.log(`Stale PID ${existing} detected (not Chrome on port ${PORT}) — launching fresh.`);
     try { unlinkSync(PID_FILE); } catch {}
@@ -152,15 +152,15 @@ async function main() {
   proc.unref();
   writeFileSync(PID_FILE, String(proc.pid));
 
-  // Wait for Chrome HTTP endpoint and build the dedicated DevToolsActivePort file
-  const portFileReady = await writePortFile();
-  if (!portFileReady) {
-    console.error('Chrome did not become ready within 15s.');
-    process.exit(1);
-  }
-
-  console.log(`Ready. No more "Allow remote debugging?" dialogs.`);
-  console.log('GreedySearch now uses its own isolated DevToolsActivePort file.');
-}
+  // Wait for Chrome HTTP endpoint and build the dedicated DevToolsActivePort file
+  const portFileReady = await writePortFile();
+  if (!portFileReady) {
+    console.error('Chrome did not become ready within 15s.');
+    process.exit(1);
+  }
+
+  console.log(`Ready. No more "Allow remote debugging?" dialogs.`);
+  console.log('GreedySearch now uses its own isolated DevToolsActivePort file.');
+}
 
 main();

package/newfeaturesideas.md

@@ -0,0 +1,105 @@
+# New Feature Ideas
+
+Ideas for future features — thinking from the perspective of an AI assistant using these tools.
+
+---
+
+## 1. Source Verification
+
+**Problem:** I get sources but can't verify if they're live, updated, or actually support the claimed content.
+
+```
+verify_sources({ urls: ["https://...", "https://..."] })
+→ [{ url, status: 200, title, snippet, lastModified, claim: "supports X" }]
+```
+
+**Use cases:**
+- Before citing a source, verify it's not 404
+- Check if a page actually contains the claimed information
+- Get last-modified dates to assess freshness
+
+---
+
+## 2. Incremental / Continuation Research
+
+**Problem:** After deep_research on "RAG vs fine-tuning", going deeper on just RAG means re-running everything with a new query and losing original context.
+
+```
+deep_research({ query: "RAG vs fine-tuning", ... })  // initial
+continue_research({ previousId: "...", query: "production RAG architectures" })  // goes deeper on RAG
+```
+
+**Use cases:**
+- Drill into a specific aspect after initial broad research
+- Build on previous results without re-fetching everything
+- Progressive disclosure of complex topics
+
+---
+
+## 3. Multi-Query Synthesis
+
+**Problem:** One query isn't enough for complex research. I chain multiple greedy_search calls manually.
+
+```
+multi_research({
+  queries: ["auth best practices", "NextAuth vs Clerk vs Lucia", "Next.js auth security"],
+  synthesize: true
+})
+```
+
+**Use cases:**
+- "Best auth for Next.js" needs multiple angles
+- Research with different facets (comparison, security, performance)
+- Casting a wider net when single query returns narrow results
+
+---
+
+## 4. Structured Extraction
+
+**Problem:** When researching "which libraries are maintained", I want tables (name, stars, last commit, license), not prose.
+
+```
+extract_structured({
+  query: "Python HTTP client libraries 2026",
+  schema: { name: "string", stars: "number", lastUpdated: "date", async: "boolean" }
+})
+```
+
+**Use cases:**
+- Library comparisons as structured data
+- Dependency audits
+- Feature matrices for tools/frameworks
+
+---
+
+## 5. Confidence Scoring on Specific Claims
+
+**Problem:** I say "high confidence" but it's hand-wavy. What if I could ask: "how confident are we that library X is actively maintained?"
+
+```
+verify_claim({
+  claim: "Prisma is actively maintained",
+  evidence: ["last commit: 2 weeks ago", "open issues: 45", "npm downloads: 2M/week"]
+})
+→ { confidence: 0.95, reasoning: "..." }
+```
+
+---
+
+## 6. Research Cache / History
+
+**Problem:** I do expensive deep_research, then the user asks a follow-up. I have to re-run everything.
+
+```
+get_research(id: "...")  // retrieve previous results
+list_research({ query: "RAG" })  // find related previous research
+```
+
+---
+
+## Priority
+
+1. **Source verification** — high value, relatively simple, fixes trust gap
+2. **Multi-query synthesis** — high value, complex but powerful
+3. **Incremental research** — medium value, nice UX improvement
+4. **Structured extraction** — medium value, specialized use cases

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@apmantza/greedysearch-pi",
-  "version": "1.4.1",
+  "version": "1.4.2",
   "description": "Pi extension: browser-automation tool that searches Perplexity, Bing Copilot, and Google AI in parallel, extracts answers and sources via CDP, with optional Gemini synthesis — grounded AI answers from real browser interactions.",
   "type": "module",
   "keywords": [

package/skills/greedy-search/SKILL.md

@@ -1,145 +1,145 @@
----
-name: greedy-search
-description: Multi-engine AI web search — greedy_search, deep_research, and coding_task. Use for high-quality research where training data may be stale or single-engine results are insufficient.
----
-
-# GreedySearch Tools
-
-## Tool Overview
-
-| Tool | Speed | Use for |
-|------|-------|---------|
-| `greedy_search` | 15-90s | Quick lookups, comparisons, debugging errors |
-| `deep_research` | 60-120s | Architecture decisions, thorough research, source-backed answers |
-| `coding_task` | 60-180s | Second opinions on code, reviews, debugging tricky issues |
-
-## When to Use Which
-
-- **`greedy_search`** — Default. Fast enough for most things. Use when you need current info.
-- **`deep_research`** — When the answer *matters*. Gives you a structured document with confidence scores, deduplicated sources ranked by consensus, Gemini synthesis, AND actual content from top sources.
-- **`coding_task`** — When you need a "second opinion" on hard problems. Best for `debug` and `plan` modes on tricky issues.
-
----
-
-# greedy_search
-
-Multi-engine AI web search with streaming progress.
-
-```greedy_search({ query: "what changed in React 19", engine: "all" })```
-
-| Parameter | Type | Default | Description |
-|-----------|------|---------|-------------|
-| `query` | string | required | The search question |
-| `engine` | string | `"all"` | `all`, `perplexity`, `bing`, `google`, `gemini` |
-| `synthesize` | boolean | `false` | Synthesize via Gemini |
-| `fullAnswer` | boolean | `false` | Complete answer vs ~300 char summary |
-
-**When to use:** Quick lookups, error messages, comparing tools, "what's new in X".
-
----
-
-# deep_research
-
-Comprehensive research with source fetching and synthesis. Returns a structured document.
-
-```deep_research({ query: "RAG vs fine-tuning for production" })```
-
-Returns:
-- Full answers from all 3 engines (Perplexity, Bing, Google)
-- Gemini synthesis combining all perspectives
-- Deduplicated sources ranked by consensus (3/3 > 2/3 > 1/3)
-- Fetched content from top 5 sources (no CDP — uses native fetch)
-- Confidence metadata (which engines responded, consensus score)
-
-**When to use:** Architecture decisions, "which library should I use", research for a writeup, anything where you need source-backed confidence.
-
----
-
-# coding_task
-
-Browser-based coding assistant using Gemini and/or Copilot.
-
-```coding_task({ task: "debug this race condition", mode: "debug", engine: "all" })```
-
-| Parameter | Type | Default | Description |
-|-----------|------|---------|-------------|
-| `task` | string | required | The coding task/question |
-| `engine` | string | `"gemini"` | `gemini`, `copilot`, or `all` |
-| `mode` | string | `"code"` | See modes below |
-| `context` | string | — | Code snippet to include |
-
-**Modes:**
-
-| Mode | Use when |
-|------|----------|
-| `debug` | Stuck on a tricky bug. Fresh eyes catch different failure modes. |
-| `plan` | About to refactor something big. Gemini plays devil's advocate. |
-| `review` | Code review before merge. High-stakes code benefits from second opinion. |
-| `test` | Need edge cases the author missed. |
-| `code` | Just need the code written (but you can probably do this yourself faster). |
-
-**When to use:** Debugging tricky issues, planning major refactors, security-critical reviews. **Skip for** simple code generation — you're faster.
-
-## Greedy Search vs Built-in Web Search
-
-| | `web_search` | `greedy_search` |
-|---|---|---|
-| Speed | Instant (~2s) | 15-60s (one engine) / 30-90s (all engines) |
-| Quality | Good for simple lookups | Higher — 3 AI engines cross-verify |
-| Synthesis | Single engine answer | Optional Gemini synthesis (cleanest answer) |
-| Use for | Quick facts, simple questions | Research, decisions, complex topics |
-
-**Rule of thumb:** Use `web_search` for quick facts. Use `greedy_search` when the answer matters — architecture decisions, comparing libraries, understanding new releases, debugging tricky errors.
-
-## When to Use
-
-- **Version-specific changes** — "What changed in React 19?" / "Breaking changes in FastAPI 0.100"
-- **Choosing between tools** — "Prisma vs Drizzle in 2026" / "Best auth library for Next.js 15"
-- **Debugging** — User pastes an error message or stack trace
-- **Research tasks** — When you need to synthesize information from multiple sources
-- **Best practices** — "How to structure a monorepo" / "Auth patterns for SaaS"
-- **Anything where training data might be stale** — 2025+, 2026+, "latest", "current", "still maintained"
-
-## Engine Selection
-
-```greedy_search({ query: "what changed in React 19", engine: "all" })```
-
-| Engine | Latency | Best for |
-|---|---|---|
-| `all` (default) | 30-90s | Highest confidence — all 3 engines in parallel |
-| `perplexity` | 15-30s | Technical Q&A, code explanations, documentation |
-| `bing` | 15-30s | Recent news, Microsoft ecosystem |
-| `google` | 15-30s | Broad coverage, multiple perspectives |
-| `gemini` | 15-30s | Google's perspective, different training data |
-
-Use a single engine when speed matters and the question isn't contentious.
-
-## Synthesis Mode
-
-For complex research questions, use `synthesize: true` with `engine: "all"`:
-
-```greedy_search({ query: "best auth patterns for SaaS in 2026", engine: "all", synthesize: true })```
-
-This deduplicates sources across engines and feeds them to Gemini for one clean, synthesized answer. Adds ~30s but produces the highest quality output — ideal for research tasks where you'd otherwise need to parse 3 separate answers.
-
-Use synthesis when:
-- You need one definitive answer, not multiple perspectives
-- You're researching a topic to write about or make a decision
-- The question has a lot of noise and you want the signal
-
-Skip synthesis when:
-- You want to see where engines disagree (useful for controversial topics)
-- Speed matters
-
-## Full vs Short Answers
-
-Default mode returns ~300 char summaries to save tokens. Use `fullAnswer: true` when you need the complete response:
-
-```greedy_search({ query: "explain the React compiler", engine: "perplexity", fullAnswer: true })```
-
-## Interpreting Results
-
-- **All 3 agree** → High confidence, present as fact
-- **2 agree, 1 differs** → Likely correct but note the dissent
-- **All differ** → Present the different perspectives to the user
-- **Sources with `[3/3]` or `[2/3]`** → Cited by multiple engines, higher confidence
+---
+name: greedy-search
+description: Multi-engine AI web search — greedy_search, deep_research, and coding_task. Use for high-quality research where training data may be stale or single-engine results are insufficient.
+---
+
+# GreedySearch Tools
+
+## Tool Overview
+
+| Tool | Speed | Use for |
+|------|-------|---------|
+| `greedy_search` | 15-90s | Quick lookups, comparisons, debugging errors |
+| `deep_research` | 60-120s | Architecture decisions, thorough research, source-backed answers |
+| `coding_task` | 60-180s | Second opinions on code, reviews, debugging tricky issues |
+
+## When to Use Which
+
+- **`greedy_search`** — Default. Fast enough for most things. Use when you need current info.
+- **`deep_research`** — When the answer *matters*. Gives you a structured document with confidence scores, deduplicated sources ranked by consensus, Gemini synthesis, AND actual content from top sources.
+- **`coding_task`** — When you need a "second opinion" on hard problems. Best for `debug` and `plan` modes on tricky issues.
+
+---
+
+# greedy_search
+
+Multi-engine AI web search with streaming progress.
+
+```greedy_search({ query: "what changed in React 19", engine: "all" })```
+
+| Parameter | Type | Default | Description |
+|-----------|------|---------|-------------|
+| `query` | string | required | The search question |
+| `engine` | string | `"all"` | `all`, `perplexity`, `bing`, `google`, `gemini` |
+| `synthesize` | boolean | `false` | Synthesize via Gemini |
+| `fullAnswer` | boolean | `false` | Complete answer vs ~300 char summary |
+
+**When to use:** Quick lookups, error messages, comparing tools, "what's new in X".
+
+---
+
+# deep_research
+
+Comprehensive research with source fetching and synthesis. Returns a structured document.
+
+```deep_research({ query: "RAG vs fine-tuning for production" })```
+
+Returns:
+- Full answers from all 3 engines (Perplexity, Bing, Google)
+- Gemini synthesis combining all perspectives
+- Deduplicated sources ranked by consensus (3/3 > 2/3 > 1/3)
+- Fetched content from top 5 sources (no CDP — uses native fetch)
+- Confidence metadata (which engines responded, consensus score)
+
+**When to use:** Architecture decisions, "which library should I use", research for a writeup, anything where you need source-backed confidence.
+
+---
+
+# coding_task
+
+Browser-based coding assistant using Gemini and/or Copilot.
+
+```coding_task({ task: "debug this race condition", mode: "debug", engine: "all" })```
+
+| Parameter | Type | Default | Description |
+|-----------|------|---------|-------------|
+| `task` | string | required | The coding task/question |
+| `engine` | string | `"gemini"` | `gemini`, `copilot`, or `all` |
+| `mode` | string | `"code"` | See modes below |
+| `context` | string | — | Code snippet to include |
+
+**Modes:**
+
+| Mode | Use when |
+|------|----------|
+| `debug` | Stuck on a tricky bug. Fresh eyes catch different failure modes. |
+| `plan` | About to refactor something big. Gemini plays devil's advocate. |
+| `review` | Code review before merge. High-stakes code benefits from second opinion. |
+| `test` | Need edge cases the author missed. |
+| `code` | Just need the code written (but you can probably do this yourself faster). |
+
+**When to use:** Debugging tricky issues, planning major refactors, security-critical reviews. **Skip for** simple code generation — you're faster.
+
+## Greedy Search vs Built-in Web Search
+
+| | `web_search` | `greedy_search` |
+|---|---|---|
+| Speed | Instant (~2s) | 15-60s (one engine) / 30-90s (all engines) |
+| Quality | Good for simple lookups | Higher — 3 AI engines cross-verify |
+| Synthesis | Single engine answer | Optional Gemini synthesis (cleanest answer) |
+| Use for | Quick facts, simple questions | Research, decisions, complex topics |
+
+**Rule of thumb:** Use `web_search` for quick facts. Use `greedy_search` when the answer matters — architecture decisions, comparing libraries, understanding new releases, debugging tricky errors.
+
+## When to Use
+
+- **Version-specific changes** — "What changed in React 19?" / "Breaking changes in FastAPI 0.100"
+- **Choosing between tools** — "Prisma vs Drizzle in 2026" / "Best auth library for Next.js 15"
+- **Debugging** — User pastes an error message or stack trace
+- **Research tasks** — When you need to synthesize information from multiple sources
+- **Best practices** — "How to structure a monorepo" / "Auth patterns for SaaS"
+- **Anything where training data might be stale** — 2025+, 2026+, "latest", "current", "still maintained"
+
+## Engine Selection
+
+```greedy_search({ query: "what changed in React 19", engine: "all" })```
+
+| Engine | Latency | Best for |
+|---|---|---|
+| `all` (default) | 30-90s | Highest confidence — all 3 engines in parallel |
+| `perplexity` | 15-30s | Technical Q&A, code explanations, documentation |
+| `bing` | 15-30s | Recent news, Microsoft ecosystem |
+| `google` | 15-30s | Broad coverage, multiple perspectives |
+| `gemini` | 15-30s | Google's perspective, different training data |
+
+Use a single engine when speed matters and the question isn't contentious.
+
+## Synthesis Mode
+
+For complex research questions, use `synthesize: true` with `engine: "all"`:
+
+```greedy_search({ query: "best auth patterns for SaaS in 2026", engine: "all", synthesize: true })```
+
+This deduplicates sources across engines and feeds them to Gemini for one clean, synthesized answer. Adds ~30s but produces the highest quality output — ideal for research tasks where you'd otherwise need to parse 3 separate answers.
+
+Use synthesis when:
+- You need one definitive answer, not multiple perspectives
+- You're researching a topic to write about or make a decision
+- The question has a lot of noise and you want the signal
+
+Skip synthesis when:
+- You want to see where engines disagree (useful for controversial topics)
+- Speed matters
+
+## Full vs Short Answers
+
+Default mode returns ~300 char summaries to save tokens. Use `fullAnswer: true` when you need the complete response:
+
+```greedy_search({ query: "explain the React compiler", engine: "perplexity", fullAnswer: true })```
+
+## Interpreting Results
+
+- **All 3 agree** → High confidence, present as fact
+- **2 agree, 1 differs** → Likely correct but note the dissent
+- **All differ** → Present the different perspectives to the user
+- **Sources with `[3/3]` or `[2/3]`** → Cited by multiple engines, higher confidence