agentsclimarketplace

Market research synthesis

Skill SylphxAI/skills/skills/market-research-synthesis

Produce a source-bounded market synthesis for a product, positioning, category, competitor, pricing, or demand decision. Use when current external evidence must be collected, triangulated, challenged with counterevidence, and converted into an original recommendation. Do not use for summarizing one supplied source, maintaining a research repository, or executing a decision already made.From its SKILL.md

Install
npx -y skills add SylphxAI/skills --skill market-research-synthesis

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

  • 24 days oldThe repository was created 24 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
  • 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

4.2 KB, 755 tokens by cl100k_base, as published. Nobody here has run it

Market Research Synthesis

Produce a decision-ready synthesis whose claims can be traced to current sources. The artifact is not a competitor scrapbook or a collection of unsupported market opinions.

Workflow

  1. Define the decision, research question, target user/buyer, category, geography, time window, product form, and evidence that could change the decision.
  2. Read references/research-synthesis-method.md. Create the source ledger before collecting conclusions. When the decision is product positioning, also read references/positioning-decision.md and bind every positioning component to that ledger.
  3. Select leaders, challengers, niche winners, substitutes, and relevant failed or low-trust examples. Explain why each belongs in the comparison.
  4. Prefer primary surfaces and independently corroborate material claims where possible. Record direct observation, attributed user report, inference, and judgment as different evidence classes.
  5. Search deliberately for counterevidence: contradictory cohorts, failed entrants, different geographies or plans, switching costs, selection bias, and facts that weaken the preferred story.
  6. Compare promises, users, first value, pricing/package, workflow, distribution, complaints, trust gaps, and defensibility at the same scope and date.
  7. Separate table stakes, copied conventions, genuine differentiators, and unresolved assumptions. Link every recommendation to supporting and contradicting ledger entries.
  8. Deliver an original wedge, product or positioning choices, risks, confidence, and the smallest next validation that could overturn the recommendation.

Source verification

  • Record URL or source identifier, publisher, access date, geography, plan, currency, evidence class, direct excerpt or observation, and confidence.
  • Triangulate consequential claims with a primary source plus an independent source when available; say when only one source exists.
  • Preserve conflicts instead of averaging them away.
  • If source access is unavailable, return a research plan and mark market claims unverified rather than filling gaps from memory.

When not to use

  • Use evidence-synthesis for a reproducible cross-domain systematic, rapid, scoping, or structured evidence review whose artifact is not a market recommendation.
  • Use source-to-skill-distiller when one bounded document or corpus must become a reusable procedure.
  • Use saas-subscription-pricing when current market evidence already exists and the primary artifact is the actual package and price decision.
  • Do not use for a routine implementation plan or a repository of raw research notes with no current decision.

Guardrails

  • Do not copy competitor wording, screenshots, proprietary workflows, or other protected expression into the recommendation.
  • Do not treat marketing claims, scraped reviews, community anecdotes, or search rank as equivalent to verified behavior or demand.
  • Do not count multiple pages repeating one claim as independent corroboration.
  • Do not hide negative evidence, source conflicts, weak coverage, or uncertainty.
  • Do not recommend dark patterns merely because competitors use them.

Output format

Decision and scope:

Source ledger:

ClaimSourceAccessedMarket/planEvidence classSupports/contradictsConfidence

Competitor and substitute comparison:

Table stakes, conventions, and differentiation:

Counterevidence and unresolved conflicts:

Recommendation, risks, confidence, and disconfirming test:

Positioning decision when requested:

  • segment/job, alternative, category, differentiator/outcome, proof, fit boundary
  • positioning statement, message hierarchy, objections and channel handoff

Keep looking

Skills are one crate of 326,736. 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.