feat: restore security validation and add unified download tools

CLAUDE.md

@@ -7,7 +7,8 @@
 Scientify 是一个 OpenClaw 插件，提供 AI 驱动的科研工作流自动化功能。
 
 **核心组件：**
-- `src/tools/` - 工具实现（arxiv_search, arxiv_download, openalex_search, unpaywall_download, github_search, paper_browser）
+- `src/tools/` - 工具实现（arxiv_search, openalex_search, paper_download, paper_browser）
+- `src/utils/` - 通用工具（security.ts 输入校验）
 - `src/commands.ts` - 聊天命令处理
 - `skills/` - 技能定义（随 npm 包发布）
 - `index.ts` - 插件入口
@@ -119,19 +120,30 @@ scientify/
 ├── README.zh.md             # 中文文档
 ├── src/
 │   ├── commands.ts          # 聊天命令
-│   ├── openclaw.d.ts        # 类型声明
+│   ├── types.ts             # 类型定义
+│   ├── cli/
+│   │   └── research.ts      # CLI 入口（init/delete/list 项目）
+│   ├── commands/
+│   │   └── metabolism-status.ts  # 新陈代谢状态命令
+│   ├── hooks/
+│   │   ├── inject-skill.ts       # Skill 注入 hook
+│   │   └── cron-skill-inject.ts  # Cron skill 注入 hook
+│   ├── templates/
+│   │   └── bootstrap.ts     # 项目初始化模板
+│   ├── utils/
+│   │   └── security.ts      # 输入校验（路径遍历防护、ID 格式校验）
 │   └── tools/
 │       ├── arxiv-search.ts         # ArXiv 搜索工具
-│       ├── arxiv-download.ts       # ArXiv 下载工具（含速率限制）
 │       ├── openalex-search.ts      # OpenAlex 跨学科搜索
-│       ├── unpaywall-download.ts   # Unpaywall OA PDF 下载
-│       ├── github-search-tool.ts   # GitHub 搜索工具
-│       └── paper-browser.ts        # 论文分页浏览工具
+│       ├── paper-download.ts       # 论文下载（arXiv .tex/PDF + DOI via Unpaywall）
+│       ├── paper-browser.ts        # 论文分页浏览工具
+│       └── result.ts               # 工具结果辅助函数
 ├── skills/
 │   ├── idea-generation/
 │   │   ├── SKILL.md
 │   │   └── references/idea-template.md
-│   ├── research-collect/SKILL.md  # 文献搜索 → 筛选 → 下载 → 聚类
+│   ├── metabolism/SKILL.md         # 新陈代谢循环（Day 0 + Day 1+）
+│   ├── research-collect/SKILL.md   # 文献搜索 → 筛选 → 下载 → 聚类
 │   ├── research-pipeline/SKILL.md  # 编排器，通过 sessions_spawn 调度以下 5 个 skill
 │   ├── research-survey/SKILL.md    # 深度论文分析 + 方法对比
 │   ├── research-plan/SKILL.md      # 四部分实现计划

README.en.md

@@ -90,7 +90,7 @@ Driven by multi-agent iteration: the orchestrator holds hypotheses and all accum
 │                          │──→│  edit them too               │
 │  arxiv_search            │   └──────────────────────────────┘
 │  openalex_search         │
-│  github_search           │
+│  paper_download          │
 │  paper_browser           │
 │  code_executor           │
 └──────────────────────────┘
@@ -121,7 +121,7 @@ The agents' hands and eyes:
 | Tool | Capability |
 |------|-----------|
 | `arxiv_search` / `openalex_search` | Search academic papers (arXiv + cross-disciplinary) |
-| `github_search` | Search open-source code implementations |
+| `paper_download` | Download papers by arXiv ID (.tex source with PDF fallback) or DOI (open-access PDF via Unpaywall). Input validation prevents shell injection. |
 | `paper_browser` | Paginated paper reading, avoids context overflow |
 | `code_executor` | Execute experiment code in `uv`-isolated environment |
 
@@ -276,7 +276,6 @@ Check status anytime:
 | Skill | Description |
 |-------|-------------|
 | **write-review-paper** | Draft a review/survey paper from project research outputs. |
-| **research-subscription** | Create/list/remove scheduled Scientify jobs via `scientify_cron_job` (research digests or plain reminders). |
 
 </details>
 
@@ -286,14 +285,9 @@ Check status anytime:
 | Tool | Description |
 |------|-------------|
 | `arxiv_search` | Search arXiv papers. Returns metadata (title, authors, abstract, ID). Supports sorting by relevance/date and date filtering. |
-| `arxiv_download` | Batch download papers by arXiv ID. Prefers .tex source files (PDF fallback). |
 | `openalex_search` | Search cross-disciplinary academic papers via OpenAlex API. Returns DOI, authors, citation count, OA status. |
-| `openreview_lookup` | Lookup OpenReview evidence by title/ID/forum. Returns decision, review rating/confidence aggregates, and review summaries. |
-| `unpaywall_download` | Download open access PDFs by DOI via Unpaywall API. Non-OA papers are silently skipped. |
-| `github_search` | Search GitHub repositories. Returns repo name, description, stars, URL. Supports language filtering and sorting. |
+| `paper_download` | Download papers by arXiv ID (prefers .tex source, PDF fallback) or DOI (open-access PDF via Unpaywall). Validates all identifiers to prevent injection. Max 20 papers per call with rate limiting. |
 | `paper_browser` | Paginated browsing of large paper files (.tex/.md) to avoid context overflow. |
-| `scientify_cron_job` | Manage scheduled Scientify jobs (`upsert`/`list`/`remove`). |
-| `scientify_literature_state` | Persistent incremental state for subscriptions: dedupe, record, feedback, and status inspection. |
 
 </details>
 
@@ -308,9 +302,6 @@ Check status anytime:
 | `/projects` | List all projects |
 | `/project-switch <id>` | Switch active project |
 | `/project-delete <id>` | Delete a project |
-| `/research-subscribe ...` | Create/update scheduled Scientify jobs |
-| `/research-subscriptions` | Show your scheduled Scientify jobs |
-| `/research-unsubscribe [job-id]` | Remove your scheduled Scientify jobs |
 
 </details>
 

README.md

@@ -90,7 +90,7 @@ Scientify 采用**新陈代谢模式**——持续地摄入、消化、沉淀、
 │                          │──→│                              │
 │  arxiv_search            │   └──────────────────────────────┘
 │  openalex_search         │
-│  github_search           │
+│  paper_download          │
 │  paper_browser           │
 │  code_executor           │
 └──────────────────────────┘
@@ -121,7 +121,7 @@ Agent 的手和眼：
 | 工具 | 能力 |
 |------|------|
 | `arxiv_search` / `openalex_search` | 搜索学术论文（arXiv + 跨学科） |
-| `github_search` | 搜索开源代码实现 |
+| `paper_download` | 下载论文：arXiv ID（优先 .tex 源文件，PDF 兜底）或 DOI（通过 Unpaywall 获取 OA PDF）。内置输入校验防止注入攻击。 |
 | `paper_browser` | 分页精读论文，避免上下文溢出 |
 | `code_executor` | 在 `uv` 隔离环境中执行实验代码 |
 

index.ts

@@ -10,6 +10,8 @@ import {
 } from "./src/commands.js";
 import { createArxivSearchTool } from "./src/tools/arxiv-search.js";
 import { createOpenAlexSearchTool } from "./src/tools/openalex-search.js";
+import { createPaperDownloadTool } from "./src/tools/paper-download.js";
+import { createPaperBrowserTool } from "./src/tools/paper-browser.js";
 import { createSkillInjectionHook } from "./src/hooks/inject-skill.js";
 import { createCronSkillInjectionHook } from "./src/hooks/cron-skill-inject.js";
 import { registerResearchCli } from "./src/cli/research.js";
@@ -23,6 +25,8 @@ export default definePluginEntry({
     // Register tools
     api.registerTool(createArxivSearchTool());
     api.registerTool(createOpenAlexSearchTool());
+    api.registerTool(createPaperDownloadTool());
+    api.registerTool(createPaperBrowserTool());
 
     // Register chat commands (bypass LLM)
     api.registerCommand({

openclaw.plugin.json

@@ -16,7 +16,6 @@
     "skills/research-review",
     "skills/research-experiment",
     "skills/research-collect",
-    "skills/paper-download",
     "skills/write-review-paper",
     "skills/metabolism"
   ]

skills/idea-generation/SKILL.md

@@ -90,7 +90,7 @@ arxiv_search({ query: "{user_topic}", max_results: 10 })
 openalex_search({ query: "{user_topic}", max_results: 10 })
 ```
 
-2. **Download papers:** 按 /paper-download 的方式下载到 `papers/`
+2. **Download papers:** 使用 `paper_download` 工具下载到 `papers/`
 
 3. **Clone reference repos (optional):**
 ```bash

skills/metabolism/SKILL.md

@@ -96,7 +96,7 @@ openalex_search({
 
 合并结果，按 arXiv ID / DOI 去重，**跳过 `processed_ids` 中已有的论文**。
 
-按 /paper-download 的方式下载新论文到 `papers/`（arXiv 优先 .tex 源文件，DOI 通过 Unpaywall 获取 OA PDF）。
+使用 `paper_download` 工具下载新论文到 `papers/`（arXiv 优先 .tex 源文件，DOI 通过 Unpaywall 获取 OA PDF）。
 
 ### Step 2: Read（阅读）
 

skills/research-collect/SKILL.md

@@ -61,7 +61,7 @@ openalex_search({ query: "<term>", max_results: 20 })
 
 #### 2.3 下载论文
 
-按 /paper-download 的方式下载论文到 `papers/`。
+使用 `paper_download` 工具下载论文到 `papers/`。
 
 **完成一个检索词后，再进行下一个。** 这样避免上下文被大量搜索结果污染。
 
@@ -140,5 +140,5 @@ mv "papers/2401.12345" "papers/data-driven/"
 |----------------|---------|
 | `arxiv_search` | 搜索 arXiv 论文 |
 | `openalex_search` | 搜索跨学科论文（覆盖更广） |
-| /paper-download | 下载论文（arXiv .tex/PDF、DOI via Unpaywall） |
+| paper_download | 下载论文（arXiv .tex/PDF、DOI via Unpaywall） |
 | `gh search repos "query"` | 搜索 GitHub 仓库 |

src/cli/research.ts

@@ -2,6 +2,7 @@ import fs from "node:fs";
 import path from "node:path";
 import os from "node:os";
 import { renderBootstrapMd, renderSoulMd, renderAgentsMd } from "../templates/bootstrap.js";
+import { validateProjectId, ensureWithinDirectory } from "../utils/security.js";
 
 const OPENCLAW_HOME = path.join(os.homedir(), ".openclaw");
 const OPENCLAW_CONFIG = path.join(OPENCLAW_HOME, "openclaw.json");
@@ -141,8 +142,10 @@ function listResearchProjects(): ResearchProject[] {
 }
 
 function initProject(id: string, pluginSkillsDir: string): void {
+  validateProjectId(id);
   const agentId = `research-${id}`;
   const workspace = path.join(OPENCLAW_HOME, `workspace-${agentId}`);
+  ensureWithinDirectory(workspace, OPENCLAW_HOME);
 
   // Check if workspace already exists
   if (fs.existsSync(workspace)) {
@@ -244,8 +247,10 @@ function showStatus(id: string): void {
 }
 
 function deleteProject(id: string): void {
+  validateProjectId(id);
   const agentId = `research-${id}`;
   const workspace = path.join(OPENCLAW_HOME, `workspace-${agentId}`);
+  ensureWithinDirectory(workspace, OPENCLAW_HOME);
 
   // Remove workspace
   if (fs.existsSync(workspace)) {

src/commands.ts

@@ -2,6 +2,7 @@ import * as fs from "node:fs";
 import * as path from "node:path";
 import * as os from "node:os";
 import type { PluginCommandContext, PluginCommandResult } from "./types.js";
+import { ensureWithinDirectory } from "./utils/security.js";
 
 const OPENCLAW_HOME = path.join(os.homedir(), ".openclaw");
 
@@ -20,10 +21,11 @@ function listResearchAgents(): ResearchAgent[] {
     const agents = (config.agents as { list?: Array<{ id: string; workspace?: string }> })?.list ?? [];
     return agents
       .filter((a) => a.id.startsWith("research-"))
-      .map((a) => ({
-        id: a.id,
-        workspace: (a.workspace ?? `~/.openclaw/workspace-${a.id}`).replace("~", os.homedir()),
-      }));
+      .map((a) => {
+        const workspace = (a.workspace ?? `~/.openclaw/workspace-${a.id}`).replace("~", os.homedir());
+        ensureWithinDirectory(workspace, OPENCLAW_HOME);
+        return { id: a.id, workspace };
+      });
   } catch {
     return [];
   }