kolega-skills

Research complex questions that require synthesis of current or distributed evidence from multiple external or user-supplied sources, comparison of competing claims, fact-checking, or an evidence-backed conclusion, and produce citation-grounded Markdown reports. Use for in-depth cultural, product, market, policy, scientific, technical, historical, community, regulatory, or experiential research. Do not use for single-fact lookups, ordinary result lists, one-URL summaries without validation, codebase-only investigation, citation formatting, or prose editing without new research.

Create, inspect, edit, and convert PowerPoint presentations (`.pptx`) with template-first layouts, structured text, images, image extraction, editable tables and charts, speaker notes, hyperlinks, and guarded slide operations. Use for presentation-native work such as building a deck from an outline or template, inventorying slide content, replacing text or media, updating tables/charts/notes, adding or removing table rows and columns, setting or removing hyperlinks, adding/removing/reordering slides, or converting PPTX to PDF. Do not use for PDF-native editing or for spreadsheets and word-processing documents.

Create, inspect, edit, extract text from, convert, and render Microsoft Word DOCX documents while preserving formatting and validating OOXML safely. Use for .docx authoring, reading, structured edits, templates or letterheads, tables, images, headers, footers, fields, DOCX-to-text, DOCX-to-PDF, and page-rendering-for-review requests. Do not use for PDF-native, spreadsheet, or presentation work.

Create, inspect, render, extract, edit, secure, redact, reorganize, and convert PDF-native documents; diagnose scanned or hybrid PDFs, run explicitly selected local OCR, and compose searchable PDFs from OCR results. Use for PDF page operations, forms, metadata, encryption, spatial text or table extraction, image/text-to-PDF generation, page rasterization for visual review, true content redaction, scanned-PDF OCR planning, and making scanned PDFs searchable. Do not use for authoring DOCX, XLSX, or PPTX source files before PDF export.

Create, inspect, extract, edit, clean, summarize, convert, and render Microsoft Excel .xlsx workbooks, including formulas, styles, tables, conditional formatting, validation, native charts, CSV/TSV interchange, PDF export, and PNG page rendering for visual review. Use for spreadsheet artifact work where workbook semantics, safe tabular conversion, or reviewable spreadsheet output matter; do not use for legacy .xls, macro-enabled .xlsm, PDF-native editing, or general analysis that does not require a spreadsheet artifact.

Create, revise, review, and package standards-compliant Agent Skills with effective trigger descriptions, progressive disclosure, reusable resources, and focused validation. Use when an agent needs to design a new skill, improve an existing SKILL.md, organize scripts/references/assets, or audit a skill before distribution.

Review GitHub pull requests, local uncommitted changes, or the current branch against a base branch for concrete correctness, security, concurrency, data-loss, compatibility, and scoped AGENTS.md violations. Use for high-signal code-review requests, `/review help`, optional GitHub inline publication, and merge-risk analysis; do not use to review prose, summarize changes without defect analysis, or implement existing review feedback.

{{description}}