Evaluate PRD section drafts against four rubrics and output verdict, score, and triage decisions.
npx skills add https://github.com/wlee075/chatbot --skill reflection-wlee075
Please help me install this Agent Skill. Skill: reflection Source: https://github.com/wlee075/chatbot/tree/main/skills/reflection Command: npx skills add https://github.com/wlee075/chatbot --skill reflection-wlee075
Reflection provides a structured, automated gate to quality by evaluating drafts against four rubrics and producing a formal verdict, triage, and recovery signals to guide section advancement.
Invoke the reflector on the current draft and PRD sections to generate a structured evaluation.
High-intent search queries and answers about installing and using this skill.
Draft quality assessment uses a four-rubric scoring system for completeness, specificity, consistency, and implementability to generate an overall score, formal verdict, and triage decision.
Triage routing for PRD sections uses recovery counters generated from the quality assessment to govern workflow progression, determining whether a draft advances or requires revision based on identified requirement gaps.
Automating quality gates for draft narratives requires invoking a reflector on the current draft to parse outputs including overall score, verdict, triage decisions, and a detailed reflective summary for reviewers.
Generating a structured verdict requires REFLECTOR_SYSTEM prompts and reflect_node integration to parse the draft and output the formal quality assessment, overall score, and identified requirement gaps.
Rubric scoring applies to both PRD sections and draft narratives, evaluating them against the four rubrics to produce a structured quality assessment, triage decisions, and recovery signals for section advancement.
Automated quality assessment limitations include dependency on REFLECTOR_SYSTEM prompts and reflect_node integration to parse outputs, meaning without these environment prerequisites, verdict and triage generation will not function.
sunchenddGenerates formatted patent drafts with claims from technical descriptions.
LumenWipeIdentify weaknesses, assumptions, and risks in generated drafts through structured elicitation methods.
IgorGanapolskyAggregate news, social, and market sentiment into composite scores for trading decisions.
tyevansCategorizes input items into predefined rubric-based verdicts, emitting grouped results with bolded labels.
