Paper AIGC Reducer (baibaiAIGC)

What it does

Paper AIGC Reducer provides a disciplined, script-backed two-round rewriting flow to reduce AI-like phrasing in academic and technical documents. For Chinese texts it enforces two sequential rounds of prompts (baibaiaigc1 -> baibaiaigc2); for English it runs a single-pass prompt. The skill coordinates reading/writing of round records and intermediate files so multi-session workflows can continue across conversations.

When to use it

Use when you need to humanize or ‘de-AIGC’ academic papers, theses, technical reports, or other scholarly text without changing facts, terminology, or logical structure. Ideal for: authors prepping submissions, editors polishing drafts, or platforms offering staged rewrite pipelines. Avoid when new facts, citations, or content expansion are required.

What's included

• Scripts: integrates with the repo's scripts (skill_round_helper.py, aigc_round_service.py, run_aigc_round.py, docx_pipeline.py) to perform chunking, round orchestration, and record updates.
• References: prompt files for each round (prompts/baibaiaigc1.md, prompts/baibaiaigc2.md, prompts/baibaiaigc-en.md) and aigc_records.json manifest conventions.
• Instructions: explicit constraints (no inventing facts, preserve terminology, single-round-per-invocation, chunk-by-chunk processing) and record-maintenance guidance.

Compatible agents

Best used with agents that can call scripts or execute managed multi-step workflows (Code-enabled agents, CLI-driven automations, or server-side orchestrators).