ai-researchlisted

External evidence with citations via a 4-tier escalation (local → free MCPs → web/Exa → NotebookLM autonomous deep research, launched first / harvested last). Every claim sourced [N] or marked [unsourced]; output ends with 3 cited recommended directions. Trigger for 'what does the state of the art say about', 'compare options for', 'find sources on', 'investigate this question', 'research this'. Use for questions whose answer lives OUTSIDE the codebase. Not for codebase exploration; use /ai-explore instead. Not for refactors; use /ai-simplify instead. Not for business-logic debugging; use /ai-debug instead.
arcasilesgroup/ai-engineering · ★ 49 · AI & Automation · score 84

Install: claude install-skill arcasilesgroup/ai-engineering

# Research ## Purpose Multi-tier, multi-source research skill with citation-first synthesis and persistent artifact reuse. Replaces ad-hoc `WebSearch` invocations with a disciplined escalation, but with NotebookLM's autonomous deep research running **async-first**: when NotebookLM is available it is launched FIRST (at T0, in a background subagent) so its slow autonomous deep-research job overlaps the fast tiers, then it is harvested LAST with a bounded wait. The fast tiers run meanwhile -- local context first (zero cost), then free MCPs (Context7, Microsoft Learn, `gh search`), then web search (Exa, with a built-in `WebSearch` fallback). Every external claim carries a `[N]` citation or is marked `[unsourced]` so readers can audit grounding, and the output always ends with exactly 3 recommended strategic directions. Every external tier (NotebookLM, Context7, Exa, MS Learn) is capability-detected and fail-soft: an absent or unauthenticated tool is skipped silently (recorded as degraded) and never errors the run. Outputs are designed for reuse: research is persisted to `.ai-engineering/runtime/research/<topic-slug>-<YYYY-MM-DD>.md` -- including the NotebookLM deep-research report and `notebook_id` -- so subsequent sessions short-circuit at Tier 0 or harvest a still-running notebook via `--reuse-notebook`. ## When to Use - User asks for evidence: "what does the industry do for X", "state of the art on Y", "compare A vs B", "find sources on Z". - `/ai-brainstorm` interrogati