What problem does it solve? Writing an academic paper involves many disconnected steps—research, drafting, citation verification, peer review, revision, and formatting—and errors like hallucinated citations or unverified claims easily slip through when these steps are done ad hoc. This Skill coordinates the entire research-to-publication pipeline with mandatory integrity gates and user confirmation checkpoints so no stage is skipped. ## Core Features & Use Cases - 10-Stage Orchestration: Dispatches deep-research, academic-paper, and academic-paper-reviewer skills across research, writing, integrity verification, two-stage peer review, revision, final integrity check, formatting, and process summary stages. - Blocking Integrity Gates: Stage 2.5 and Stage 4.5 run 100% reference/citation verification plus a 7-mode AI research failure checklist (implementation bugs, hallucinated results, methodology fabrication, and more) that blocks the pipeline until resolved. - Adaptive Checkpoints & State Tracking: FULL/SLIM/MANDATORY checkpoints require user confirmation, a Material Passport enables cross-session resume via resume_from_passport=<hash>, and an optional claim-faithfulness audit agent verifies claims against retrieved source text. - Use Case: A researcher says "I want to write a paper on AI in higher education quality assurance" and the pipeline runs from literature research through drafting, integrity verification, simulated peer review, revision, and final PDF/DOCX output with a bilingual process record. ## Quick Start Ask the assistant to run the full academic pipeline for your research topic, for example by saying you want to write a research paper on a given topic from scratch.
