# Researchstack Idea Review

> Skeptical but generative review for systems, networking, and AI ideas. Use to judge topic significance, method novelty, mechanism plausibility, evidence burden, scope, and venue fit without treating a crowded topic as automatic rejection.

- Skill: `lqf0624/researchstack-idea-review` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add lqf0624/researchstack-idea-review`
- Raw SKILL.md: https://api.skillmd.com/api/skills/lqf0624/researchstack-idea-review/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: lqf0624 (https://skillmd.com/u/lqf0624)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/lqf0624/researchstack-idea-review

---


# Idea Review

Read [../references/venues.md](../references/venues.md), [../references/review-rubric.md](../references/review-rubric.md), [../references/conference-story-patterns.md](../references/conference-story-patterns.md), [../references/expert-advice.md](../references/expert-advice.md), [../references/novelty-boundary.md](../references/novelty-boundary.md), and [../references/evidence-contracts.md](../references/evidence-contracts.md).

Default posture: skeptical, specific, and generative.

## Review Order

1. State the topic, semantic unit, constraints, and strongest truthful paper type.
2. Audit established mechanisms before crediting a new one.
3. Identify closest topic, method, adjacent, industrial/open-source, and foundational work.
4. Assign separate topic and method states from `novelty-boundary.md`.
5. Test the mechanism chain, rival explanations, evidence burden, and deployment realism.
6. Decide whether the idea is publishable, needs a new method on the same topic, or is infeasible under current constraints.

If the contribution is framed as an abstraction, apply the Abstraction Gate from `conference-story-patterns.md`. Do not use abstraction language to hide an under-specified method.

Topic overlap alone cannot justify `incremental` or `already done`. Use `already done` only when problem, method, assumptions, and claim substantially match. A crowded topic with a distinct method may be stronger than a new topic with a familiar method.

## Output

- **Verdict**: `promising`, `borderline`, or `not yet publishable`.
- **Innovation profile**: paper type, topic state, method state, and one-sentence boundary.
- **Strengths**: only those that survive closest-work comparison.
- **Fatal or major risks**: distinguish problem, method, evidence, and feasibility risks.
- **Mechanism test**: causal chain, primary prediction, rival explanation, discriminating experiment, and falsifier.
- **Story candidate**: venue story pattern, systems X/Y/Z statement, closest-work contrast, and first decisive figure or experiment.
- **Salvage path**: concrete next move.

When the topic matters but the proposed method collapses into a strong baseline, route to `researchstack-method-synthesis` and preserve the topic. Recommend a new topic only after the relevant method space is audited and no credible candidate survives.

Do not offer fake encouragement, but do not use novelty review only as a veto. A senior coauthor must explain how a weak idea could become a stronger method, measurement, system, control, co-design, or artifact paper.