Rebuttal

wanshuiyin/auto-claude-code-research-in-sleep

When external reviews land and you need to write a conference rebuttal, this workflow parses the feedback, atomizes every concern into a tracked issue board, and drafts a venue-compliant response with strict provenance gates. It supports both single-document rebuttals (ICML style) and per-reviewer threads (OpenReview), enforces character limits, and stress-tests your draft against a simulated reviewer before you submit. The safety model is rigid: every claim needs a source (paper, review, or user-confirmed result), every promise needs approval, and every issue gets explicitly answered or deferred. It won't run new experiments or fabricate results, but if you bring approved evidence it will weave it in. Quick mode stops after strategy planning if you just want to see what reviewers are asking for.

skills.sh installs: 532
GitHub stars: 18.9K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add wanshuiyin/auto-claude-code-research-in-sleep --skill rebuttal --agent claude-code

Installs into .claude/skills of the current project.

Files

Workflow 4: Rebuttal

Before each reviewer slot, follow Selectable Reviewers: resolve invocation → .aris/reviewer.json → Claude Code default (Codex MCP) with reviewer_route.py through the canonical helper resolver, preserving this slot's tier, fresh contexts, source/image inputs and deterministic gates. Codex examples/pins below apply only to backend=codex; other routes use their own first-call/reply contract. Save the route and actual returned identity; unavailable selections are REVIEW_UNAVAILABLE. Classify acceptance from actual families/completion/scope. Forward invocation reviewer/model/effort overrides to reviewer-bearing children. Generation workers are unchanged.

Prepare and maintain a grounded, venue-compliant rebuttal for: $ARGUMENTS