Problem Identification

name problem-identification version 0.1.0 description Identifies and articulates the core problem, motivation, and gap that a technical document (such as a research paper or systems article) is addressing. tags ["documents","problem","research"] Purpose This skill is used to extract a precise description of the main problem or question the document addresses, why it matters, and what gap in prior work or practice it claims to fill. The goal is to provide a clear, compact statement of: What issue the authors are trying to solve or understand. Why this issue is important. How existing approaches or systems are insufficient. Expected Input The complete document text as a single string. Optionally, a sections structure produced by the structure-understanding skill that identifies major sections (e.g., abstract, introduction, background). The agent using this skill is responsible for: Obtaining the document text (e.g., via a PDF extraction tool), and When available, obtaining the sections structure first. When a sections structure is available, this skill should primarily focus on sections whose role is "abstract" , "introduction" , "background" , or any other role that clearly introduces motivation and context. Expected Output Return a JSON-like structure with the following shape: { "problem_statement" : "One or two sentences that clearly state the main problem or question the document addresses." , "motivation" : "Two to four sentences explaining why this problem matters in practice or research, including any real-world impact or bottlenecks." , "gap_in_prior_work" : "One to three sentences describing what prior methods, systems, or understanding are missing or inadequate, as claimed by the document." } If the document does not explicitly discuss one of these elements, infer only what is strongly supported by the text. If no clear information is available for a field, set it to a short statement such as: "The document does not explicitly state this." Behavioral Guidelines When a sections structure is available: Start from sections labeled "abstract" , "introduction" , and "background" . Prefer content that explicitly states goals, challenges, bottlenecks, or research questions. Use the section summaries as guidance but base your final wording on the full text when needed. Focus on what the problem is, not how it is solved (methods belong to other skills). Avoid repeating generic phrases; be specific to the document’s domain and context. Do not copy sentences verbatim from the original text. Paraphrase in clear, direct language. If the document describes several sub-problems, merge them into a coherent main problem statement and explain secondary aspects in the motivation or gap fields. Example (informal) For a document that introduces a framework to reduce the latency of AI agents that interact with environments through sequential API calls, a valid output might look like: { "problem_statement" : "The document addresses the problem that AI agents interacting with complex environments via sequential API calls suffer from high end-to-end latency, making them impractical for interactive use and large-scale training." , "motivation" : "As agents are increasingly deployed in environments such as browsers, operating systems, games, and e-commerce systems, the time they spend waiting on each tool call or API response dominates their overall runtime. This delays user interactions, slows experimentation, and makes iterative optimization expensive." , "gap_in_prior_work" : "Existing work on speculative decoding focuses mainly on speeding up token-level generation within a single LLM, and does not generalize to the full agent-environment loop. There is a lack of a general, lossless framework for speculating on future actions across diverse agentic environments." }