advisor

Analyzes research papers and proposals to judge composition, currency, feasibility, and next experiments.

2|2|Updated Jun 21, 2026
One-click install
npx skills add https://github.com/BTreeMap/SKILLs --skill advisor-btreemap
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: advisor
Source: https://github.com/BTreeMap/SKILLs/tree/main/advisor
Command: npx skills add https://github.com/BTreeMap/SKILLs --skill advisor-btreemap

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Researchers and lab leads need a rigorous, principal-investigator-style read of a paper, proposal, or result set: what prior ideas it combines, whether its experimental setup matches current practice, what the lab can actually afford, and which cheap experiment would prove or kill each claim. ## Core Features & Use Cases - Four-move analysis: Decomposes a project into its constitution (origins tagged as-is, tweaked, transferred, or new), currency of each setup element (current, legacy, or toy, with dated sources), the lab's resource envelope, and per-claim instruments with kill tests. - Five verbs: review (default, read-only judgment of one artifact), design (phased experiment plan with costs and kill tests), audit (program-level portfolio judgment), teach (audience-calibrated explanation citing sourced heuristics), and help (quick-reference card). - Use Case: Share a draft systems paper and ask whether the setup holds up; the skill reads the method section first, classes each setup element against current field practice with retrieved sources, and names the cheapest credible experiment per claim. ## Quick Start Ask the advisor to review the attached paper draft and tell you what it is made of, whether the setup is current, and what experiment to run next.

Frequently Asked Questions about advisor

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get a critical review of a research paper draft?

Invoke the review verb with the paper attached. The skill reads the method section and citations first, decomposes the work into its prior-idea origins, classes each setup element as current, legacy, or toy with sources, and gives a direction paragraph naming the first claim to test.

How to plan the next experiments for a research project?

Use the design verb after a review. It produces a phased plan where phase one is the cheapest experiment that can kill a claim, each phase carries a cost in the lab's envelope, a kill test, and the decision its result unlocks.

Can it judge a whole research program or lab portfolio?

Yes, the audit verb samples projects by leverage, runs the four-move analysis on each sampled project, names the unsampled ones with reasons, and recommends which projects to merge, stop, or run first.

Does the skill verify claims about current research practice?

Every currency claim must carry a dated source retrieved during the run, or be explicitly marked as from memory with the model's cutoff. Retrieval uses the harness's search and fetch tools or the search-web skill.

When should I use peer-review instead of this advisor?

Use peer-review when you want an adverse referee report pressing on claims, design, analysis, and novelty. The advisor instead reads a project as its principal investigator, focusing on composition, feasibility, and the cheapest decisive next experiment.