agentsclimarketplace

Investgame research

Skill InvestGameLtd/investgame-skills/skills/investgame-research

Use for DEEP work - when the user wants more than a quick answer: a report, a market or regional landscape, a deep dive, a trend analysis, a company/deal write-up, an investor profile, prep for a call or meeting, OR turning a call / meeting transcript into structured notes. Triggers: "report on / state of / landscape of [market]", "deep dive / deep research on [topic]", "research [company/deal/trend]", "profile [company or investor]", "prepare me for a call/meeting with [company]", "call / meeting notes", "make a deck / PDF / PowerPoint about [topic]", or any broad, comprehensive task. Most tasks run a plan-first workflow (investigate, propose a plan, then build a benchmarked, on-brand deliverable using investgame-gaming-data, investgame-analysis and investgame-format, plus web research); a transcript runs a transcript-processing workflow instead. NOT for a single data lookup (answer that with the investgame-gaming-data hub). Not for game-design or generic software questions.From its SKILL.md

Install
npx -y skills add InvestGameLtd/investgame-skills --skill investgame-research

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

7.8 KB, as published. Nobody here has run it

InvestGame Research & Reports

The deep-work layer. The everyday skills (data via the investgame-gaming-data hub, analysis, format) answer a question; this one handles the bigger ask - a report, a landscape, a deep dive, company or trend research, an investor profile, prep for a meeting. Whatever the shape, the job is the same: investigate comprehensively, prepare the material properly, and deliver it on-brand. A "report" is just the most common instance of this.

When this activates

A request that needs depth, not a single number: "a market report on …", "the state of … gaming", "a deep dive / deep research on …", "research this company / deal / trend", "profile this investor", "prepare me for a meeting with …", "make a deck / PDF about …", or any broad, comprehensive task. A quick fact is NOT this -answer that directly with the investgame-gaming-data hub.

Plan first - always

Deep work starts with a plan, not with output. Never jump straight to a 15-slide deck.

  1. Understand the task - what question, which subject (a market / segment / company / investor / trend), how deep, and for what audience.
  2. Investigate enough to plan well - a quick first pass over the data and the sources (below) so the plan is grounded, not guessed.
  3. Propose a short, human-readable plan - a table of contents: the sections/slides with a sub-bullet each ("Slide 1 - market at a glance; Slide 2 - deal activity & trend; …"). Aim for ~5 sections for a focused ask, up to ~15 for a broad one. Ask which theme (Warm White or Dark Navy). Invite changes.
  4. On approval (or if the user already said "just do it"), execute. Don't build the full thing before the plan is agreed. Keep the plan concise - a TOC with light sub-bullets, not an essay.

Gather comprehensively (two streams, run them together)

A deep deliverable needs a rounded picture, so pull from more than one angle.

  • InvestGame data - via the investgame-gaming-data hub. Don't pull one cut; assemble several so the story is complete. Standard cuts (unless the user narrows them): top ~15 deals; top ~15 buyers / most-active investors / most-funded companies; and historical dynamics over time across dimensions - value (deal size), count (number of deals/events), and valuation (multiples). Full recipes in references/playbook.md.
  • Research - the narrative and outside perspective a report needs. Prioritise sources in order:
    1. investgame.net first - the digest and features pages already cover the deals and themes; start there.
    2. the curated gaming-industry sources in references/research-sources.md (analysts, newsletters, market data) - prefer these over generic search results.
    3. general web - for broader context and narrative. (Web research uses your own web tools; InvestGame's data tool doesn't browse.) If your agent has no web/browsing tool, deliver the InvestGame data layers in full and clearly LABEL the narrative, market-size and sourcing sections as gaps for the user to fill (or for InvestGame custom research) - never fabricate the outside-context stream from memory. Web research is a capability that may be absent; say so rather than quietly dropping or inventing it.

Read, benchmark, present

  • Read & benchmark via investgame-analysis - never describe a market or company in isolation; put it in context (peers, the global total, prior years). The comparison is the value. Analysis's rules carry over (multiples, never-break-silently, footnote assumptions, objective-not-opinionated, the critical-thinking pass before you finish).
  • Present via investgame-format - on-brand, in the theme the user picked. Format owns the look and the form: a deck/PDF for a full report, a one-pager or a written brief for lighter prep. For market-size figures InvestGame doesn't hold, cite third-party sources separately, average and show the range, and never blend them into InvestGame deal numbers.

What you can produce (instances of the same workflow)

  • A market / regional / segment report - activity, sizing, leaders, outlook, benchmarked.
  • A deep dive / trend research - a theme over time, grounded in the data.
  • A company or deal research write-up; an investor / fund profile.
  • A meeting or call preparation brief - a focused, deep profile to walk in ready.
  • Structured notes / a recap from a call or meeting transcript - topic-organised, decision-focused, with single-owner action items, enriched with InvestGame context (see the notes mode below). See references/playbook.md for the section recipes and the standard data cuts per type.

Notes from a call or meeting transcript (a different mode - no plan step)

When the input is a transcript rather than a research question, don't plan-first - process it:

  1. Clarify only if needed - unidentified speakers, missing company/context, ambiguous action-item ownership, or poor transcript quality (many [inaudible]). Match the output language to the transcript.
  2. Process - map speakers to names/roles; group by topic, not chronologically; extract decisions (what was decided, not just discussed); extract action items (task · single owner · deadline · priority); note open questions.
  3. Enrich via the hub - for any gaming company or deal mentioned, pull concise InvestGame context (recent funding, last deal, valuation, segment) into an "InvestGame context" block; you may add a quick web check for very recent news. Keep it factual; flag anything unverified.
  4. Render via investgame-format - a clean, branded notes document. Structure: participants · subject · summary (≤6 bullets of decisions/outcomes, not topics) · discussion by topic · InvestGame context · action items table · open questions · next steps. Call notes (short 1:1 / intro / diligence) stay tight; meeting notes (multi-party) are fuller - same logic, different weight. If unsure, ask "quick call recap or full meeting record?"

Charter (non-negotiable)

Objective, never sell-side - present what the data and sources show, not how to "win" a pitch. People and companies described neutrally. Honest sourcing: label every source, flag gaps ("insufficient public data") instead of inventing, no hype.

Common mistakes

  • Jumping to a full deck before the plan is agreed.
  • Pulling one data cut - a deep report needs several angles (top deals, top players, the time trend).
  • Skipping investgame.net and the curated sources, going straight to generic web.
  • Describing in isolation instead of benchmarking.
  • Blending third-party market-size figures into InvestGame deal numbers without labelling.
  • Treating a quick lookup as deep work - just answer it with the investgame-gaming-data hub.
  • For a transcript: transcribing chronologically instead of synthesising by topic, or listing discussion topics as if they were decisions.

Keep looking

Skills are one crate of 608,475. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.