research

Investigate questions against primary sources and write numbered research artefact files.

3|Updated Jan 10, 2026
One-click install
npx skills add https://github.com/MaxWolf-01/agents --skill research-maxwolf-01
Or copy as Structured Prompt for Agentβ–Ό
Please help me install this Agent Skill.
Skill: research
Source: https://github.com/MaxWolf-01/agents/tree/main/mx/skills/research
Command: npx skills add https://github.com/MaxWolf-01/agents --skill research-maxwolf-01

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Agent sessions often make decisions on unverified assumptions or secondhand information, and findings get lost between sessions. This Skill investigates a question against primary sources and persists the findings as a numbered research artefact that future sessions can consume. ## Core Features & Use Cases - Frame checking: Detects when a question arrives with an inherited premise or option list, states the shared frame, and tests the strongest option outside it before investigating. - Primary-source investigation: Follows every claim back to official docs, source code, specs, or first-party APIs rather than secondary write-ups. - Structured artefacts: Writes findings to agent/research/NN-<slug>.md with Question, Frame, and Findings sections, each claim cited by URL, file path, or commit, and cross-linked to related tickets via frontmatter. - Use Case: A ticket proposes three caching options. Run the research skill to test the framing, verify each option against the library's official documentation, and produce a cited artefact the implementation session reads directly. ## Quick Start Ask the agent to research whether the current pagination approach in the API client handles cursor expiry correctly and write up the findings.

Frequently Asked Questions about research

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

FAQPage Schema
How do I research a technical question with an AI agent?β–Ό

Invoke the research skill with a topic or question as its argument. The agent sharpens fuzzy questions, investigates against primary sources like official docs and source code, and writes a cited artefact to agent/research/ with Question, Frame, and Findings sections.

What is a research artefact in an agent workflow?β–Ό

A research artefact is a numbered Markdown file (agent/research/NN-slug.md) containing point-in-time findings that other sessions consume. It cites every claim by URL, file path, or commit and links back to related tickets through frontmatter.

How does frame checking work in research tasks?β–Ό

When a question arrives with a diagnosis or option list attached, the skill states the shared premise in one sentence, names the strongest option outside that frame, and checks whether dropping the premise dissolves the problem. User-decided frames like ADRs stand; conflicts are surfaced, not reopened.

Why should research cite primary sources instead of blog posts?β–Ό

Secondary write-ups can be outdated, simplified, or wrong, and agents that rely on them propagate errors into specs and code. Following each claim back to official docs, source code, specs, or first-party APIs ensures findings rest on the source that owns them.

When should research findings be linked to a ticket?β–Ό

Link the artefact whenever the research backs a spec or ticket, since future agents read from the ticket direction. List the artefact on the ticket and add a ticket: frontmatter line in the artefact pointing back.