What problem does it solve? When users encounter errors in the cannbot-knowledge knowledge base—such as incorrect content, retrieval failures, graph errors, or downstream agent issues—they often submit incomplete bug reports that cannot be acted upon. This Skill structures user feedback into complete, reproducible GitCode issue submissions with all required evidence. ## Core Features & Use Cases - Structured Issue Generation: Produces a complete 16-section issue template covering problem type, versions, environment, prompts, commands, terminal output, reproduction rate, and evidence. - Material Validation: Checks 14 required materials and outputs a needs-info list when critical items like version commits, full terminal logs, or reproduction records are missing. - Reproduction Package Assembly: Defines a standardized attachment archive structure (environment.txt, command.txt, prompt.txt, terminal.log, sessions/, reproduction runs) for triage-ready submissions. - Use Case: A user finds that a knowledge query returns wrong results for an Ascend operator. The Skill collects the knowledge base commit, CLI version, verbatim query, three reproduction runs, and affected doc-id, then outputs a ready-to-file issue plus a packaged evidence archive. ## Quick Start Help me file a knowledge base issue: the knowledge-query search for 'A5 LayerNorm UB reuse' returns an incorrect document, and here are my command, terminal output, and repo commit.
