Creates factual QRDSI research artifacts documenting system behavior, constraints, code paths, vocabulary, risks and unknowns.
npx skills add https://github.com/luan/agents --skill research-luan
Please help me install this Agent Skill. Skill: research Source: https://github.com/luan/agents/tree/main/skills/research Command: npx skills add https://github.com/luan/agents --skill research-luan
It prevents design and implementation from being driven by assumptions by creating a factual QRDSPI research artifact about how the system behaves today.
Create or update a research artifact by asking: research “what is the current behavior and code path for <topic>?” --depth high --auto.
High-intent search queries and answers about installing and using this skill.
To document current system behavior before a refactor, you need an evidence-first research artifact that captures verified facts, constraints, and code paths rather than proposed solutions. This prevents design and implementation from being driven by assumptions.
The best way to gather verified evidence for system research is by exploring the codebase and vault to capture behavior, constraints, vocabulary, risks, and unknowns, ensuring the artifact accurately describes current system reality.
To preserve existing context when drafting a research artifact, load relevant existing artifacts from the vault and carry forward resolved user decisions, defaults, non-goals, and scope boundaries to avoid relitigating previously settled topics.
Yes, you can obtain artifact approval before committing by running a Plannotator gate for the absolute path, ensuring the research artifact describing the system reality is fully accepted before any commit occurs.
You need to produce a factual research artifact for pre-design or pre-implementation phases when evidence is required before planning, structure, or implementation decisions can be safely made without relying on assumptions.
qmuGenerate structured release notes from branch story frontmatter to Markdown.
luanCreate a QRDSPI design artifact with tradeoffs and review gating.
arestivoExplore codebases and clarify requirements without implementing changes.
GiorgioRicciardielloAnalyze tournament debates and reflection reviews to synthesize patterns into a structured meta-review note.
