Paper Summarize Pdf To Feishu

总结论文 PDF 为飞书文档（含图表）。采用主控 - 子代理 (Orchestrator-Subagents) 架构处理长流程。 支持学术论文、技术报告的自动去重、总结、配图、真实比对审核与人工确认。 触发场景：发送 PDF 附件、"总结这个 PDF"、"把 PDF 写成飞书文档"、"论文总结"。

This skill mostly does what it says (extract text, OCR figures, build summaries, check duplicates) and all logic is in included shell scripts — that's good. However, before installing or running: 1) Confirm how Feishu publishing is authenticated: the skill references feishu_doc actions and feishu_doc_token.txt but declares no Feishu credential; do not place API tokens in global or shared directories. 2) Define and lock down $PAPER_DIR and $PAPERS_DIR to a dedicated sandbox/workspace so the scripts cannot traverse your entire papers directory or other sensitive folders. 3) Expect the agent to spawn subagents and write logs; review the logs and output location. 4) If you want to publish to Feishu, require the maintainer to declare a clear primaryEnv (e.g., FEISHU_API_TOKEN) and document where/how tokens are supplied. 5) If you cannot verify how feishu_doc actions are executed (what service/tool actually performs network requests), run the skill offline or in a sandbox first and inspect produced commands and any network calls. If these issues are not clarified, treat the skill as risky for production use.

Type: OpenClaw Skill Name: paper-summarize-pdf-to-feishu Version: 2.0.0 The skill implements a complex multi-agent workflow for summarizing PDF papers into Feishu documents, but it exhibits high-risk behaviors. Most notably, in SKILL.md, it instructs the agent to read the sensitive global configuration file `~/.openclaw/openclaw.json` (which typically contains API keys) to programmatically identify vision-capable models. Additionally, it requires `sudo` privileges to install several system dependencies and executes multiple shell scripts (e.g., `locate_figures.sh`, `extract_metadata.sh`) to process files. While these actions are plausibly related to the stated functionality, accessing the platform's core configuration file is a significant security risk that could be leveraged for credential theft.