researchlisted

<!-- Vendored from mattpocock/skills (MIT), upstream commit 2ab9580 (2026-07-28), plugin v1.2.0, adapted for n8n-project-template. See .claude/skills/_LICENSE-mattpocock.md --> Spin up a **background agent** to do the research, so you keep working while it reads. If you cannot spawn agents in your current context (e.g. you are a pipeline sub-agent), run the research inline instead — the rules below apply either way. Its job: 1. Investigate the question against **primary sources** — official docs, source code, specs, first-party APIs — not a secondary write-up of them. Follow every claim back to the source that owns it. 2. Write the findings to a single Markdown file, citing each claim's source — **every factual claim carries its source URL plus the retrieval date**. 3. Save it where this repo keeps such notes: service-specific integration research goes to `docs/integrations/<service>/` (precedent: `docs/integrations/m365/`); everything else goes to `docs/research/<topic>.md` (create the directory if it does not exist). ## Research backends Use them in this order of preference: 1. **`mcp__perplexity__perplexity_research`** — deep multi-source investigation. If the tool is deferred, load its schema via ToolSearch first. 2. **Context7** (`mcp__context7__resolve-library-id` + `query-docs`) — library, framework, and API documentation. 3. **WebFetch / WebSearch** as fallback — pointed at official primary sources only. Whichever backend answers, the primary-source rule above