productivity▌

Remove backgrounds from images via inference.sh CLI.

Create interactive HTML flowcharts and process diagrams.

Design visually impactful presentation slides with layout concepts, visual recommendations, and design specifications. \n \n Provides 8+ reusable slide layout patterns including title slides, data highlights, comparisons, timelines, and process flows with ASCII wireframes \n Generates complete design specifications covering layout placement, color palettes, typography guidance, icon recommendations, and animation suggestions \n Offers design direction templates (minimalist, corporate, creative,

Iterative evaluation and refinement patterns for improving AI agent outputs through self-critique loops. \n \n Provides three core patterns: basic reflection (self-critique loops), evaluator-optimizer (separated generation and evaluation), and code-specific test-driven refinement \n Supports multiple evaluation strategies including outcome-based assessment, LLM-as-judge comparison, and rubric-based scoring with weighted dimensions \n Includes practical Python implementations with structured JSON

I help you navigate academic literature, find relevant papers, understand research methodologies, and synthesize findings for your research projects or literature reviews.

Design high-engagement carousel posts via inference.sh CLI.

Use this skill as the semantics companion to idiomatic-zener for non-trivial .zen work.

Search and retrieve Reddit posts, comments, subreddit info, and user profiles via the public JSON API. \n \n Requires no API key; uses Reddit's public JSON endpoint accessible by appending .json to any URL \n Supports four main command categories: subreddit posts with sorting (hot, new, top, rising, controversial), cross-subreddit search, subreddit metadata, and user profiles with optional post history \n Includes flexible filtering: sort by time range (hour through all-time), set result limits,

When agent sessions generate millions of tokens of conversation history, compression becomes mandatory. The naive approach is aggressive compression to minimize tokens per request. The correct optimization target is tokens per task: total tokens consumed to complete a task, including re-fetching costs when compression loses critical information.

Extractive summarization from long documents with flexible length control and batch processing. \n \n Supports three algorithms (TextRank, LSA, frequency-based) with configurable language support \n Control summary length by ratio, sentence count, or word count; optionally preserve original sentence order \n Extract key points as bullet-point summaries alongside full-text summaries \n Batch process multiple documents or entire directories with consistent parameters \n Available as Python API or