oh-my-openidea 0.1.4 → 0.1.6

package/README.md

@@ -20,9 +20,7 @@ bunx oh-my-openidea@latest install
 Non-interactive mode with API keys:
 
 ```bash
-bunx oh-my-openidea@latest install --no-tui --openai=yes --anthropic=yes --tmux=no --skills=yes \
-  --serpapi-key=YOUR_SERPAPI_KEY \
-  --semantic-scholar-key=YOUR_S2_KEY
+bunx oh-my-openidea@latest install --no-tui --openai=yes --anthropic=no --antigravity=yes --copilot=no --zai-plan=no --chutes=no --kimi=yes --tmux=no --skills=yes --serpapi-key=YOUR_SERPAPI_KEY --semantic-scholar-key=YOUR_S2_KEY
 ```
 
 Then authenticate:
@@ -57,61 +55,205 @@ Additional guides:
 
 ### 01. Orchestrator: The Architect of Discovery
 
-| The one who sees all paths. | The Orchestrator is the lead coordinator for the whole research workflow. Give it a problem, and it decides which agents should work next, what information is still missing, and how the pieces should come together into a strong research direction. |
-| --- | --- |
-| Role: Strategic research coordination and idea synthesis |
-| Prompt: orchestrator.ts |
-| Recommended Models: kimi-for-coding/k2p5   openai/gpt-5.2-codex |
+<table>
+  <tr>
+    <td width="240" align="center" valign="top">
+      <img src="img/orchestrator.png" alt="Orchestrator" width="220" />
+    </td>
+    <td valign="top">
+      <strong>The one who sees all paths.</strong> The Orchestrator is
+      the lead coordinator for the whole research workflow. Give it a
+      problem, and it decides which agents should work next, what
+      information is still missing, and how the pieces should come
+      together into a strong research direction. In practice, it runs
+      the full pipeline: clarifies the problem, dispatches parallel
+      literature searches to <code>@surveyor</code>, hands the corpus
+      to <code>@synthesizer</code> for gap extraction, generates
+      3-5 concrete hypotheses itself, sends the strongest candidates
+      to <code>@critic</code> for novelty/feasibility review, then
+      calls <code>@architect</code> and <code>@writer</code> to turn
+      the winning idea into an executable plan and paper outline. It
+      is the only agent with full delegation authority and is
+      responsible for deciding when to branch in parallel and when to
+      keep steps sequential.
+    </td>
+  </tr>
+  <tr>
+    <td colspan="2">
+      <strong>Role:</strong> Strategic research coordination and idea synthesis<br/>
+      <strong>How it works:</strong> Multi-stage planner with explicit subagent routing, limited refinement loops, and final synthesis across all intermediate outputs<br/>
+      <strong>Best Model Traits:</strong> Strong long-horizon planning, reliable tool/subagent orchestration, stable structured reasoning, and the ability to compress many intermediate results into clear decisions<br/>
+      <strong>Prompt:</strong> <code>orchestrator.ts</code><br/>
+    </td>
+  </tr>
+</table>
 
 ---
 
 ### 02. Surveyor: The Cartographer of Knowledge
 
-| The one who maps the unknown. | The Surveyor finds and collects the most relevant papers for your topic. It helps you quickly understand what has already been done, which papers matter most, and where the current limits of the literature are. |
-| --- | --- |
-| Role: Literature search and paper retrieval |
-| Prompt: surveyor.ts |
-| Recommended Models: openai/gpt-5.1-codex-mini &nbsp; google/gemini-3-flash |
+<table>
+  <tr>
+    <td width="240" align="center" valign="top">
+      <img src="img/explorer.png" alt="Surveyor" width="220" />
+    </td>
+    <td valign="top">
+      <strong>The one who maps the unknown.</strong> The Surveyor finds
+      and collects the most relevant papers for your topic. It helps
+      you quickly understand what has already been done, which papers
+      matter most, and where the current limits of the literature are.
+      Its job is deliberately narrow and read-only: search, retrieve,
+      and organize. It runs parallel queries across arXiv, Semantic
+      Scholar, Google Scholar, and the web; prefers recent work by
+      default; separates foundational papers from recent advances; and
+      always returns traceable identifiers such as arXiv IDs or
+      Semantic Scholar IDs so later agents can verify claims instead
+      of relying on vague summaries.
+    </td>
+  </tr>
+  <tr>
+    <td colspan="2">
+      <strong>Role:</strong> Literature search and paper retrieval<br/>
+      <strong>How it works:</strong> Read-only retrieval agent that fans out across multiple academic sources, compares query variants, and builds a paper corpus with metadata for downstream use<br/>
+      <strong>Best Model Traits:</strong> Fast tool calling, high extraction precision, low hallucination rate on bibliographic details, and good breadth-first search behavior under ambiguous queries<br/>
+      <strong>Prompt:</strong> <code>surveyor.ts</code><br/>
+    </td>
+  </tr>
+</table>
 
 ---
 
 ### 03. Synthesizer: The Weaver of Understanding
 
-| The one who connects the dots. | The Synthesizer turns a pile of papers into a clear understanding of the field. It groups results, highlights patterns and disagreements, and points out open problems that can lead to new research ideas. |
-| --- | --- |
-| Role: Gap analysis and knowledge synthesis |
-| Prompt: synthesizer.ts |
-| Recommended Models: kimi-for-coding/k2p5 &nbsp; google/gemini-3-flash |
+<table>
+  <tr>
+    <td width="240" align="center" valign="top">
+      <img src="img/librarian.png" alt="Synthesizer" width="220" />
+    </td>
+    <td valign="top">
+      <strong>The one who connects the dots.</strong> The Synthesizer
+      turns a pile of papers into a clear understanding of the field.
+      It groups results, highlights patterns and disagreements, and
+      points out open problems that can lead to new research ideas.
+      Instead of merely summarizing, it builds a research landscape:
+      themes, trends, recurring limitations, and evidence-backed gaps.
+      It reads key sections of papers, cross-references claims across
+      the corpus, and distinguishes between gaps that are unaddressed
+      because they are genuinely hard and gaps that have simply been
+      overlooked. If the corpus is missing an important branch, it can
+      trigger a targeted follow-up search via <code>@surveyor</code>.
+    </td>
+  </tr>
+  <tr>
+    <td colspan="2">
+      <strong>Role:</strong> Gap analysis and knowledge synthesis<br/>
+      <strong>How it works:</strong> Corpus-level analyzer that reads paper sections, clusters ideas into themes, surfaces contradictions and blind spots, and outputs evidence-linked research gaps<br/>
+      <strong>Best Model Traits:</strong> Long-context reading, cross-document comparison, disciplined citation use, and strong abstraction ability without drifting away from source evidence<br/>
+      <strong>Prompt:</strong> <code>synthesizer.ts</code><br/>
+    </td>
+  </tr>
+</table>
 
 ---
 
 ### 04. Critic: The Guardian of Truth
 
-| The voice of rigorous doubt. | The Critic stress-tests your idea before you invest too much in it. It looks for weak assumptions, missing baselines, overlap with prior work, and other reasons the idea might fail novelty or review standards. |
-| --- | --- |
-| Role: Novelty validation and adversarial review |
-| Prompt: critic.ts |
-| Recommended Models: openai/gpt-5.2-codex &nbsp; kimi-for-coding/k2p5 |
+<table>
+  <tr>
+    <td width="240" align="center" valign="top">
+      <img src="img/oracle.png" alt="Critic" width="220" />
+    </td>
+    <td valign="top">
+      <strong>The voice of rigorous doubt.</strong> The Critic
+      stress-tests your idea before you invest too much in it. It
+      looks for weak assumptions, missing baselines, overlap with
+      prior work, and other reasons the idea might fail novelty or
+      review standards. Operationally, it behaves like an adversarial
+      program committee reviewer: it searches for overlapping prior
+      work, reads candidate papers to verify whether the overlap is
+      superficial or real, and scores each idea on novelty,
+      feasibility, significance, clarity, and overall quality. When
+      needed, it can trigger targeted follow-up literature checks via
+      <code>@surveyor</code>, but it does not redesign the idea
+      itself; its purpose is to reject weak directions early and make
+      promising ones harder to fool yourself about.
+    </td>
+  </tr>
+  <tr>
+    <td colspan="2">
+      <strong>Role:</strong> Novelty validation and adversarial review<br/>
+      <strong>How it works:</strong> Evidence-first evaluator that combines prior-work search, paper-level overlap verification, and structured review scoring in a NeurIPS/ICML-style format<br/>
+      <strong>Best Model Traits:</strong> Skeptical reasoning, careful distinction between related and duplicate ideas, strong comparative judgment, and consistency when turning evidence into explicit scores and verdicts<br/>
+      <strong>Prompt:</strong> <code>critic.ts</code><br/>
+    </td>
+  </tr>
+</table>
 
 ---
 
 ### 05. Architect: The Builder of Methods
 
-| The one who turns vision into plan. | The Architect converts an idea into a concrete experiment plan. It defines datasets, baselines, metrics, ablations, failure cases, and resource needs so you know exactly how to test whether the idea works. |
-| --- | --- |
-| Role: Methodology and experiment design |
-| Prompt: architect.ts |
-| Recommended Models: kimi-for-coding/k2p5 &nbsp; openai/gpt-5.1-codex-mini |
+<table>
+  <tr>
+    <td width="240" align="center" valign="top">
+      <img src="img/designer.png" alt="Architect" width="220" />
+    </td>
+    <td valign="top">
+      <strong>The one who turns vision into plan.</strong> The
+      Architect converts an idea into a concrete experiment plan. It
+      defines datasets, baselines, metrics, ablations, failure cases,
+      and resource needs so you know exactly how to test whether the
+      idea works. It takes a hypothesis that has already survived
+      critique and expands it into an executable methodology: method
+      overview, baseline stack, dataset choices, statistical tests,
+      ablation plan, implementation notes, and realistic compute
+      estimates. It is a leaf agent by design, which means it focuses
+      on turning validated ideas into concrete experimental decisions
+      rather than delegating further.
+    </td>
+  </tr>
+  <tr>
+    <td colspan="2">
+      <strong>Role:</strong> Methodology and experiment design<br/>
+      <strong>How it works:</strong> Execution planner that translates validated ideas into benchmarkable experiments with concrete baselines, datasets, metrics, ablations, and compute budgets<br/>
+      <strong>Best Model Traits:</strong> High specificity, strong methodological priors, good benchmark literacy, realistic resource estimation, and low tolerance for vague implementation advice<br/>
+      <strong>Prompt:</strong> <code>architect.ts</code><br/>
+    </td>
+  </tr>
+</table>
 
 ---
 
 ### 06. Writer: The Voice of Science
 
-| The one who makes it legible to the world. | The Writer turns your research into a clear paper draft. It helps organize the story, sharpen the main message, and present the method and results in a way that is easy for readers and reviewers to follow. |
-| --- | --- |
-| Role: Research writing and paper structure |
-| Prompt: writer.ts |
-| Recommended Models: openai/gpt-5.1-codex-mini &nbsp; google/gemini-3-flash |
+<table>
+  <tr>
+    <td width="240" align="center" valign="top">
+      <img src="img/fixer.png" alt="Writer" width="220" />
+    </td>
+    <td valign="top">
+      <strong>The one who makes it legible to the world.</strong> The
+      Writer turns your research into a clear paper draft. It helps
+      organize the story, sharpen the main message, and present the
+      method and results in a way that is easy for readers and
+      reviewers to follow. Unlike the other agents, it is intentionally
+      execution-only: no external search, no tool-driven fact hunting,
+      and no subagent delegation. It assumes the research content has
+      already been established upstream and focuses on turning that
+      content into paper structure, abstract, introduction, related
+      work narrative, and section-level draft text that follows
+      mainstream ML conference conventions.
+    </td>
+  </tr>
+  <tr>
+    <td colspan="2">
+      <strong>Role:</strong> Research writing and paper structure<br/>
+      <strong>How it works:</strong> Context-conditioned writer that receives finalized idea and methodology, then produces publication-style outlines and drafts without inventing unsupported claims<br/>
+      <strong>Best Model Traits:</strong> Strong academic writing quality, good discourse organization, precise controllability, and discipline about staying within provided evidence instead of fabricating results<br/>
+      <strong>Prompt:</strong> <code>writer.ts</code><br/>
+    </td>
+  </tr>
+</table>
 
 ---
 

package/dist/cli/index.js

@@ -13736,7 +13736,7 @@ var TmuxConfigSchema = exports_external.object({
   main_pane_size: exports_external.number().min(20).max(80).default(60)
 });
 var PresetSchema = exports_external.record(exports_external.string(), AgentOverrideConfigSchema);
-var McpNameSchema = exports_external.enum(["websearch", "arxiv", "semantic_scholar", "google_scholar"]);
+var McpNameSchema = exports_external.enum(["websearch"]);
 var BackgroundTaskConfigSchema = exports_external.object({
   maxConcurrentStarts: exports_external.number().min(1).max(50).default(10)
 });
@@ -13760,10 +13760,10 @@ var PluginConfigSchema = exports_external.object({
 // src/config/agent-mcps.ts
 var DEFAULT_AGENT_MCPS = {
   orchestrator: ["websearch"],
-  surveyor: ["websearch", "arxiv", "semantic_scholar", "google_scholar"],
-  synthesizer: ["websearch", "arxiv", "semantic_scholar"],
-  critic: ["arxiv", "semantic_scholar", "google_scholar"],
-  architect: ["websearch", "arxiv"],
+  surveyor: ["websearch"],
+  synthesizer: ["websearch"],
+  critic: [],
+  architect: ["websearch"],
   writer: ["websearch"]
 };
 
@@ -13817,6 +13817,18 @@ var CUSTOM_SKILLS = [
     description: "Repository understanding and hierarchical codemap generation",
     allowedAgents: ["orchestrator"],
     sourcePath: "src/skills/cartography"
+  },
+  {
+    name: "alphaxiv-paper-lookup",
+    description: "Look up arXiv papers via AlphaXiv structured overview",
+    allowedAgents: [
+      "orchestrator",
+      "surveyor",
+      "synthesizer",
+      "critic",
+      "architect"
+    ],
+    sourcePath: "src/skills/alphaxiv-paper-lookup"
   }
 ];
 function getCustomSkillsDir() {
@@ -16253,6 +16265,8 @@ function argsToConfig(args) {
     preferredOpenCodeModel: args.opencodeFreeModel && args.opencodeFreeModel !== "auto" ? args.opencodeFreeModel : undefined,
     artificialAnalysisApiKey: args.aaKey,
     openRouterApiKey: args.openrouterKey,
+    serpApiKey: args.serpApiKey,
+    semanticScholarKey: args.semanticScholarKey,
     balanceProviderUsage: args.balancedSpend === "yes",
     hasTmux: args.tmux === "yes",
     installSkills: args.skills === "yes",
@@ -16573,9 +16587,11 @@ async function runInteractiveMode(detected, modelsOnly = false) {
       rl.close();
       return config2;
     }
-    const totalQuestions = 11;
+    const totalQuestions = 13;
     const existingAaKey = getEnv("ARTIFICIAL_ANALYSIS_API_KEY");
     const existingOpenRouterKey = getEnv("OPENROUTER_API_KEY");
+    const existingSerpApiKey = getEnv("SERPAPI_KEY");
+    const existingSemanticScholarKey = getEnv("SEMANTIC_SCHOLAR_API_KEY");
     console.log(`${BOLD}Question 1/${totalQuestions}:${RESET}`);
     const artificialAnalysisApiKey = await askOptionalApiKey(rl, "Artificial Analysis API key for better ranking signals", existingAaKey);
     if (existingAaKey && !artificialAnalysisApiKey) {
@@ -16589,6 +16605,18 @@ async function runInteractiveMode(detected, modelsOnly = false) {
     }
     console.log();
     console.log(`${BOLD}Question 3/${totalQuestions}:${RESET}`);
+    const serpApiKey = await askOptionalApiKey(rl, "SerpAPI key for Google Scholar search (enables academic Google Scholar results)", existingSerpApiKey);
+    if (existingSerpApiKey && !serpApiKey) {
+      printInfo("Using existing SERPAPI_KEY from environment.");
+    }
+    console.log();
+    console.log(`${BOLD}Question 4/${totalQuestions}:${RESET}`);
+    const semanticScholarKey = await askOptionalApiKey(rl, "Semantic Scholar API key for higher rate limits (optional, free tier works without it)", existingSemanticScholarKey);
+    if (existingSemanticScholarKey && !semanticScholarKey) {
+      printInfo("Using existing SEMANTIC_SCHOLAR_API_KEY from environment.");
+    }
+    console.log();
+    console.log(`${BOLD}Question 5/${totalQuestions}:${RESET}`);
     const useOpenCodeFree = await askYesNo(rl, "Use Opencode Free models (opencode/*)?", "yes");
     console.log();
     let availableOpenCodeFreeModels;
@@ -16619,25 +16647,25 @@ async function runInteractiveMode(detected, modelsOnly = false) {
         console.log();
       }
     }
-    console.log(`${BOLD}Question 4/${totalQuestions}:${RESET}`);
+    console.log(`${BOLD}Question 6/${totalQuestions}:${RESET}`);
     const kimi = await askYesNo(rl, "Enable Kimi provider?", detected.hasKimi ? "yes" : "no");
     console.log();
-    console.log(`${BOLD}Question 5/${totalQuestions}:${RESET}`);
+    console.log(`${BOLD}Question 7/${totalQuestions}:${RESET}`);
     const openai = await askYesNo(rl, "Enable OpenAI provider?", detected.hasOpenAI ? "yes" : "no");
     console.log();
-    console.log(`${BOLD}Question 6/${totalQuestions}:${RESET}`);
+    console.log(`${BOLD}Question 8/${totalQuestions}:${RESET}`);
     const anthropic = await askYesNo(rl, "Enable Anthropic provider?", detected.hasAnthropic ? "yes" : "no");
     console.log();
-    console.log(`${BOLD}Question 7/${totalQuestions}:${RESET}`);
+    console.log(`${BOLD}Question 9/${totalQuestions}:${RESET}`);
     const copilot = await askYesNo(rl, "Enable GitHub Copilot provider?", detected.hasCopilot ? "yes" : "no");
     console.log();
-    console.log(`${BOLD}Question 8/${totalQuestions}:${RESET}`);
+    console.log(`${BOLD}Question 10/${totalQuestions}:${RESET}`);
     const zaiPlan = await askYesNo(rl, "Enable ZAI Coding Plan provider?", detected.hasZaiPlan ? "yes" : "no");
     console.log();
-    console.log(`${BOLD}Question 9/${totalQuestions}:${RESET}`);
+    console.log(`${BOLD}Question 11/${totalQuestions}:${RESET}`);
     const antigravity = await askYesNo(rl, "Enable Antigravity (Google) provider?", detected.hasAntigravity ? "yes" : "no");
     console.log();
-    console.log(`${BOLD}Question 10/${totalQuestions}:${RESET}`);
+    console.log(`${BOLD}Question 12/${totalQuestions}:${RESET}`);
     const chutes = await askYesNo(rl, "Enable Chutes provider?", detected.hasChutes ? "yes" : "no");
     console.log();
     if (chutes === "yes") {
@@ -16661,7 +16689,7 @@ async function runInteractiveMode(detected, modelsOnly = false) {
         console.log();
       }
     }
-    console.log(`${BOLD}Question 11/${totalQuestions}:${RESET}`);
+    console.log(`${BOLD}Question 13/${totalQuestions}:${RESET}`);
     const balancedSpend = await askYesNo(rl, "Do you have subscriptions or pay per API? If yes, we will distribute assignments evenly across selected providers so your subscriptions last longer.", "no");
     console.log();
     let skills = "no";
@@ -16703,6 +16731,8 @@ async function runInteractiveMode(detected, modelsOnly = false) {
       availableChutesModels,
       artificialAnalysisApiKey,
       openRouterApiKey,
+      serpApiKey,
+      semanticScholarKey,
       balanceProviderUsage: balancedSpend === "yes",
       hasTmux: false,
       installSkills: skills === "yes",

package/dist/cli/types.d.ts

@@ -112,6 +112,8 @@ export interface InstallConfig {
     scoringEngineVersion?: ScoringEngineVersion;
     artificialAnalysisApiKey?: string;
     openRouterApiKey?: string;
+    serpApiKey?: string;
+    semanticScholarKey?: string;
     balanceProviderUsage?: boolean;
     hasTmux: boolean;
     installSkills: boolean;

package/dist/config/schema.d.ts

@@ -98,9 +98,6 @@ export declare const PresetSchema: z.ZodRecord<z.ZodString, z.ZodObject<{
 export type Preset = z.infer<typeof PresetSchema>;
 export declare const McpNameSchema: z.ZodEnum<{
     websearch: "websearch";
-    arxiv: "arxiv";
-    semantic_scholar: "semantic_scholar";
-    google_scholar: "google_scholar";
 }>;
 export type McpName = z.infer<typeof McpNameSchema>;
 export declare const BackgroundTaskConfigSchema: z.ZodObject<{