Skill UI

unslop is a powerful command-line interface (CLI) tool designed for post-processing text. It programmatically strips out common and identifiable patterns associated with AI-generated writing. Use it as a final, deterministic pass before publishing documentation, blog posts, or any content derived from large language models (LLMs), ensuring a more natural and human-like tone.

This tool functions as a post-task epilogue, designed to audit and record the entire research journey. It scans the conversation history at the end of a coding or research session to extract critical research artifacts, including decisions, experiments, dead ends, claims, heuristics, and major pivots. It systematically writes this provenance data into the structured `ara/` directory, ensuring a faithful and auditable trace of how the project truly evolved.

This skill is dedicated to handling Markdown files stored as plain text documents within Lark Drive. It provides comprehensive functions for viewing, creating, uploading, editing, patching, and comparing Markdown content. It is crucial for localized text modifications, version comparison, and content updates, while explicitly excluding cloud space management tasks like searching, permissions, or converting to online documents.

Automatically generates a professional, single-file, horizontal-swipe HTML slide deck for weekly team updates. Ideal for routine status meetings, this skill structures key information into specific sections: what was shipped, what is in flight, blocked items, key metrics, and next week's asks. Streamlines team communication and reporting processes.

Use this comprehensive skill to generate, manipulate, and export draw.io diagrams. It converts diagrams into native .drawio files and allows exporting to high-quality formats like PNG, SVG, and PDF. Crucially, it embeds the editable XML data, ensuring the resulting diagrams are professional, scalable, and easily integrated into documents.

This skill analyzes a user's recent activity history captured by the local screenpipe daemon. It detects repeated research or operational workflows, clusters these patterns, and compares them against existing skills. It then drafts new skills or suggests composition recipes to cover unaddressed workflows, helping the user formalize their operational processes.

This skill performs a post-generation, per-page review to refine the visual quality of academic talk decks (PPTX and Beamer). It systematically addresses layout drift, adjusts font scaling for projector readability, fixes common typographical errors (e.g., italic leaks, overflow), and aligns the overall visual weight of the presentation to a specified reference PDF or style preset. It is ideal for the final polish of content-rich, visually complex academic presentations.

This skill addresses vague or ambiguous user requests by initiating a structured, multi-round questioning process. Instead of making assumptions, it systematically clarifies dimensions like audience, scope, tone, and format across several rounds (up to three). This structured approach ensures the resulting output is highly accurate and requires minimal revision cycles.

This skill analyzes and reports on the token usage, expenditure, and budget for AI models (like Claude Code) by querying a local SQLite database. It is essential for users who need detailed financial accountability, tracking costs by project, tool, or date, and managing operational budgets.

This skill formalizes the process of logging a founder's decision, turning in-session deliberation into durable company memory. It utilizes a two-layer memory model to store both approved, structured decisions (which feed into future planning) and raw, verbatim transcripts (for historical reference and preserving dissent). It supports structured decision records, success criteria, kill criteria, and follow-up scheduling.

This comprehensive 12-agent pipeline guides users through the entire academic writing lifecycle, from initial research planning to final publication-ready draft. Features include literature review, structure design, full-text drafting, citation compliance (multiple formats), bilingual abstract generation, and simulated peer review. It ensures high writing quality via style calibration and advanced checks, supporting output formats like LaTeX, DOCX, and PDF.

This is a mandatory, one-time setup skill designed to interview users interactively to build a personalized knowledge base for advanced email triage. Using a structured, discipline-driven interview process, the skill collects data on email patterns, business context, and priorities, generating 7 critical structured files (e.g., email-taxonomy.md, rate-card.md). This knowledge base powers the companion 'inbox-triage' skill. Run this before initial use or whenever core business processes change significantly.