← 返回 Skills 市场
chiefsegundo

Boof

作者 chiefsegundo · GitHub ↗ · v4.0.0 · MIT-0
cross-platform ⚠ suspicious
1007
总下载
0
收藏
7
当前安装
2
版本数
在 OpenClaw 中安装
/install boof
功能描述
Convert PDFs and documents to markdown, index them locally for RAG retrieval, and analyze them token-efficiently. Use when asked to: read/analyze/summarize a...
使用说明 (SKILL.md)

Boof 🍑

Local-first document processing: PDF → markdown → RAG index → token-efficient analysis.

Documents stay local. Only relevant chunks go to the LLM. Maximum knowledge absorption, minimum token burn.

Powered by opendataloader-pdf — #1 in PDF parsing benchmarks (0.90 overall, 0.93 table accuracy). CPU-only, no GPU required.

Quick Reference

Convert + index a document

bash {SKILL_DIR}/scripts/boof.sh /path/to/document.pdf

Convert with custom collection name

bash {SKILL_DIR}/scripts/boof.sh /path/to/document.pdf --collection my-project

Query indexed content

qmd query "your question" -c collection-name

Core Workflow

  1. Boof it: Run boof.sh on a PDF. This converts it to markdown via opendataloader-pdf (local Java engine, no API, no GPU) and indexes it into QMD for semantic search.

  2. Query it: Use qmd query to retrieve only the relevant chunks. Send those chunks to the LLM — not the entire document.

  3. Analyze it: The LLM sees focused, relevant excerpts. No wasted tokens, no lost-in-the-middle problems.

When to Use Each Approach

"Analyze this specific aspect of the paper" → Boof + query (cheapest, most focused)

"Summarize this entire document" → Boof, then read the markdown section by section. Summarize each section individually, then merge summaries. See advanced-usage.md.

"Compare findings across multiple papers" → Boof all papers into one collection, then query across them.

"Find where the paper discusses X" → qmd search "X" -c collection for exact match, qmd query "X" -c collection for semantic match.

Output Location

Converted markdown files are saved to knowledge/boofed/ by default (override with --output-dir).

Setup

If boof.sh reports missing dependencies, see setup-guide.md for installation instructions (Java + opendataloader-pdf + QMD).

Environment

  • ODL_ENV — Path to opendataloader-pdf Python venv (default: ~/.openclaw/tools/odl-env)
  • QMD_BIN — Path to qmd binary (default: ~/.bun/bin/qmd)
  • BOOF_OUTPUT_DIR — Default output directory (default: ~/.openclaw/workspace/knowledge/boofed)
安全使用建议
Boof appears internally consistent with its stated purpose, but review these practical points before installing: - The script runs locally and will execute Python/Java on your machine and write markdown to the specified output directory; only run it on documents you authorize. - Installing opendataloader-pdf and QMD requires network access and will download packages and (on first run) QMD models (~1–2GB). Verify you trust the opendataloader-pdf and QMD sources (inspect their repos or package pages) before installing. - The setup uses bun to install QMD from a GitHub URL — prefer installing in isolated environments (venv, container, or VM) if processing sensitive documents. - No credentials are requested by the skill, but ensure you set a safe BOOF_OUTPUT_DIR if you do not want converted files stored under your home/workspace. - If you need higher assurance, run the boof.sh commands manually in an isolated venv and review the opendataloader-pdf/QMD behavior during first-run model downloads.
功能分析
Type: OpenClaw Skill Name: boof Version: 4.0.0 The 'boof' skill bundle provides legitimate PDF-to-Markdown conversion and RAG indexing using local tools like opendataloader-pdf and qmd. However, the script 'scripts/boof.sh' contains a code injection vulnerability where the '$INPUT_FILE' shell variable is interpolated directly into a Python script within a heredoc. This could allow arbitrary Python code execution if the agent is tasked with processing a file with a specially crafted filename (e.g., containing quotes and Python commands). While the tool's behavior aligns with its stated purpose and no intentional malice was found, this high-risk implementation flaw warrants a suspicious classification.
能力评估
✓ Purpose & Capability
The name/description (PDF→markdown→RAG) matches the included files and script. Required tools (Java, Python venv with opendataloader-pdf, and QMD) are exactly what you would expect for local conversion and local semantic indexing. No unrelated binaries or credentials are requested.
ℹ Instruction Scope
The SKILL.md and scripts instruct the agent to run a local shell script that (a) runs opendataloader-pdf inside a venv to convert the provided file, (b) indexes the resulting markdown with qmd, and (c) writes output to a local directory. This stays within the stated purpose. Note: the setup and first-run steps will download QMD models (~1–2GB) and require network access when installing packages (pip / bun / qmd); logs are filtered in the script output which reduces noise but also hides some informational lines. The skill does not reference or exfiltrate unrelated system files or environment variables.
✓ Install Mechanism
There is no automated install spec in the skill bundle; setup instructions tell the user to install Java, pip-install opendataloader-pdf into a venv, and install QMD via bun from a GitHub URL. These are standard, traceable sources (PyPI/GitHub/bun). No arbitrary binary downloads or extract-from-unknown-URLs are present in the bundle.
✓ Credentials
No secrets or cloud credentials are required. Declared environment variables (ODL_ENV, QMD_BIN, BOOF_OUTPUT_DIR) are path/configuration variables appropriate for the task. The skill does not request unrelated tokens or passwords.
✓ Persistence & Privilege
always is false and the skill does not modify other skills or system-wide configs. It writes converted files under a local output directory (default under the workspace) which is proportional to its function.
如何使用
  1. 确保已安装 OpenClaw（本地或 Docker 部署）
  2. 在对话框中输入安装命令：/install boof
  3. 安装完成后，直接呼叫该 Skill 的名称或使用 /boof 触发
  4. 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v4.0.0
Swap Marker (GPU, slow, flat text) for opendataloader-pdf (CPU-only, #1 benchmark, proper Markdown tables). Faster, lighter, better output quality. Requires Java 11+ instead of Python ML stack.
v1.0.0
Initial release
元数据
Slug boof
版本 4.0.0
许可证 MIT-0
累计安装 7
当前安装数 7
历史版本数 2
常见问题

Boof 是什么？

Convert PDFs and documents to markdown, index them locally for RAG retrieval, and analyze them token-efficiently. Use when asked to: read/analyze/summarize a... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 1007 次。

如何安装 Boof？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install boof」即可一键安装，无需额外配置。

Boof 是免费的吗？

是的，Boof 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。

Boof 支持哪些平台？

Boof 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。

谁开发了 Boof？

由 chiefsegundo（@chiefsegundo）开发并维护，当前版本 v4.0.0。

推荐 Skills
self-improving agent
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau...
⬇ 463283 · by pskoett
Skill Vetter
Security-first skill vetting for AI agents. Use before installing any skill from ClawdHub, GitHub, or other sources. Checks for red flags, permission scope, and suspicious patterns.
⬇ 259410 · by spclaudehome
Polymarket
Query Polymarket prediction markets. Check odds, find trending markets, search events, track price movements.
⬇ 232503 · by joelchance
Self-Improving + Proactive Agent
Self-reflection + Self-criticism + Self-learning + Self-organizing memory. Agent evaluates its own work, catches mistakes, and improves permanently. Use when...
⬇ 200423 · by ivangdavila
Github
Interact with GitHub using the `gh` CLI. Use `gh issue`, `gh pr`, `gh run`, and `gh api` for issues, PRs, CI runs, and advanced queries.
⬇ 191113 · by steipete
ontology
Typed knowledge graph for structured agent memory and composable skills. Use when creating/querying entities (Person, Project, Task, Event, Document), linkin...
⬇ 190156 · by oswalpalash

💬 留言讨论