Score research ideas across five dimensions with three reviewer personas and Area Chair synthesis.
npx skills add https://github.com/OpenLAIR/dr-claw --skill inno-idea-eval
Please help me install this Agent Skill. Skill: inno-idea-eval Source: https://github.com/OpenLAIR/dr-claw/tree/main/skills/inno-idea-eval Command: npx skills add https://github.com/OpenLAIR/dr-claw --skill inno-idea-eval
This Skill provides a structured, evidence-grounded evaluation of research ideas using three reviewer personas and a final Area Chair synthesis, ensuring robust assessment before proceeding to downstream development.
Provide a research idea to the Evaluator and run Steps 1 through 4 to obtain a structured Area Chair meta-review and a clear decision.
High-intent search queries and answers about installing and using this skill.
To evaluate research idea feasibility and novelty, this Skill uses three reviewer personas to independently score Clarity, Novelty, Validity, Feasibility, and Significance, followed by an Area Chair synthesis that resolves disagreements and issues a final decision.
A multi-persona meta-review is an evaluation process where three independent reviewer personas score an idea across five dimensions, and an Area Chair aggregates scores and resolves disagreements to produce a final structured decision and human-readable critique.
You generate machine-readable JSON logs by running the evaluation workflow, which outputs eval_report, eval_scores, eval_decision, and context variables designed for downstream ideation and code-survey pipeline integration.
Yes, you can run automated idea evaluation standalone when pipeline artifacts are partially missing; the evaluator grounds judgments using available prompts and task context, providing transparent caveats for any absent references.
Yes, the idea evaluator provides a structured quality gate by producing an Area Chair meta-review with a clear decision, designed specifically for integration between the ideation and code-survey stages of a research workflow.
A limitation of automated persona validation is that it grounds judgments only in provided prompts, references, and task context; if artifacts are missing, the evaluation runs standalone with caveats, potentially impacting evidence-based validity scores.
xiaoyu-opsScore research ideas across five dimensions using three reviewer personas and an Area Chair synthesis.
Mathews-TomAutomates NotebookLM CLI workflows for creating, ingesting and analyzing research notebooks and generating artifacts.
PthahnixOrchestrate literature survey, gap analysis, ideation, and review across research rounds.
seanchatmangptRun adversarial reviews on PRs, documentation, or architecture with three review levels.
