Source-Backed Research Brief: Facts, Citations + Angles

Unlike free "research this for me" prompts that invent statistics, experts and URLs, this skill separates established vs uncertain points, rates each one high/medium/low confidence, and leaves clearly marked source slots (or uses the links you supply) instead of fabricating citations.

Sample output (excerpt — topic: agent skill marketplaces for solo creators; audience: indie hacker deciding whether to list a first paid SKILL.md):

Research brief: Agent skill marketplaces for solo creators

Audience: Indie hacker deciding go / no-go this month · Depth: standard Generated: draft for human verification — not advice

2. Key points

• KP1 — Format is converging on SKILL.md. Package what you already do well. Confidence: medium
• KP5 — Buyers distrust fake proof. Invented case studies or URLs raise refund risk. Confidence: high
• KP6 — Narrow beats broad for v1. One demo-able job sells better than "does all research". Confidence: high

4. Risks and watch-outs

Risk	Severity	Mitigation / watch
Slow or zero early sales	high	Ship, then promote weekly; iterate title/tags
Payout / KYC friction	high	Start payout setup before polishing copy
Overclaiming earnings	medium	"Illustrative" language only; no guarantees

6. Source placeholders

1. [SOURCE: Chosen marketplace fee disclosure — URL TBD]

7. Suggested next step

Finish payout KYC on one marketplace, list one focused skill, share one before/after post.

You get 7 fixed sections every time: executive snapshot, 5–8 key points with confidence, implications for your audience, risk table, open questions to verify, source slots, and one next step (about 400–800 words).

Works in Claude Code, Cursor and Codex · no invented stats, quotes or URLs · v1.0, updated Oct 2026 · questions answered via Agensi messages within 24h · More Cygnix Labs tools: https://apify.com/cygnix_labs

Turn a vague research question into a structured brief with confidence labels and source placeholders. A focused skill for coding agents (Cursor, Claude Code, Codex CLI, Gemini CLI, Copilot, and other runtimes that load SKILL.md).

What you get

Give the agent a topic and an audience (optional: goal, geography, time horizon). It returns a structured Markdown brief:

1. Executive snapshot
2. Key points with confidence labels
3. Implications for that audience
4. Risks table (severity + mitigation)
5. Open questions for human follow-up
6. Source placeholders (you fill real URLs)
7. One concrete next step

Typical length: about 400–800 words — roughly 1–2 pages when printed or pasted into Notion/Docs.

Who it’s for

Founders, PMs, consultants, freelancers, and analysts who want a repeatable first draft before meetings, pitches, or build decisions — without the agent inventing citations or overconfident forecasts.

How it behaves

• Asks for missing topic/audience instead of guessing
• Separates established points from uncertain ones
• Refuses to fabricate statistics, expert quotes, or source URLs
• Uses en-GB wording and an honesty pass (no guaranteed outcomes)

What’s included: SKILL.md installable instructions; example-output.md clearly marked sample; README.md install and usage notes.

Honest limits

This skill drafts a starting brief. You (or your client) must verify facts and replace placeholder sources before decisions that matter. It is not legal, medical, or financial advice, and it does not claim any sales or income result from using or listing skills.

Pairs well with (Cygnix Labs, on Agensi):

• Keyword Research — Clusters + Intent Brief: find and cluster keywords
• Keyword Gap Analysis: keywords competitors rank for and you don't
• SERP Intent Mapper: match each keyword to the right page type
• Topical Authority Map: pillar, clusters and internal links
• SEO Content Brief (free): writer-ready brief for any keyword
• Content Refresh Audit: fix pages that are losing Google traffic
• Local SEO Keyword Pack / Local SEO Audit: for local businesses