Research Report Curator

The problem

Local research reports, industry documents, and exported web pages often arrive as mixed folders with PDFs, slides, spreadsheets, scanned files, encrypted files, and unreadable documents. Reading them one by one is slow, and it is hard to decide which files deserve priority, which need OCR, and what a management-ready reading path should look like.

How it works

The skill starts by confirming the input scope: a folder path, uploaded files, or an existing manifest. It filters out temporary artifacts such as .DS_Store and ~$* files, then builds a manifest with file name, type, size, page count, extractability, and anomaly notes.

Core capabilities include:

• scoring files by business relevance, decision value, actionability, timeliness, data density, and credibility;
• grouping them into S / must read, A / key read, B / selective read, and C / quick scan tiers;
• producing a summary, category overview, Top N reading list, and recommended route;
• briefing a single report and, when explicitly requested, copying selected files to a target folder.

Boundaries

It is intended for organization, prioritization, and summarization, not legal, investment, audit, or medical advice. Encrypted, damaged, or unreadable files should be marked for review, and high-risk actions such as moving or overwriting must require explicit confirmation.