recent-funding-rounds

When To Use

Use for recently published news coverage of startup funding rounds within a clear topic and relative time window. The skill makes one bounded news search and returns up to ten candidate articles for evidence-based deal-flow review.

This is not a structured funding database. It does not guarantee that every article describes a newly closed round, and it cannot prove a company, stage, amount, or announcement date beyond what the returned article supports. Use a dedicated structured-data capability when the user needs exhaustive coverage or database-grade fields.

Do not use this skill for investment execution, personalized financial advice, private-contact discovery, or outreach.

Workflow

1. Collect both required inputs:

   • focus: a concrete funding-news topic such as artificial intelligence, fintech Series A, or rounds over $10M. The query appends startup funding round announced, so do not repeat "startup".
   • freshness: one of pd, pw, pm, or py, representing the past 24 hours, 7 days, 31 days, or 365 days respectively.

   Do not silently choose a sector or time window. If either is absent, ask the user. Treat these as rolling publication windows, not exact calendar periods.

2. Install the vetted recipe:

   selat skill install recent-funding-rounds

3. Run the free live verification before any wallet setup or paid call:

   SELAT_ROUTER_URL=https://router.selat.ai \
     selat skill verify ~/.config/selat/skills/recent-funding-rounds \
     --focus "artificial intelligence" \
     --freshness "pw" \
     --live-probe

4. Show the user the live quote (about $0.03675 on 2026-10-04), the $0.050 step cap, and the one-call scope. The top-level maxAmount is only the per-step fallback cap, not a full-run cap; with one step they coincide. Explain that a paid application error may still charge. Wait for explicit approval of the expected cost and maximum session cap.

5. Only after approval and a spendable Gateway balance, arm a session budget no higher than the approved cap, execute once, and stop the budget after success or failure:

   selat budget start --amount <approved-cap>
   selat skill run recent-funding-rounds \
     --focus "<funding-news-focus>" \
     --freshness "<pd|pw|pm|py>" \
     --max-amount <approved-cap>
   selat budget stop

Inspect payment history before any retry. A different focus or freshness window is a new paid call and requires a fresh probe and separate approval.

Inputs And Outputs

The fixed request searches for ${focus} startup funding round announced, requests ten results, and applies the supplied publication-freshness window.

Distill the response into:

1. The supplied focus, rolling publication window, and retrieval time.
2. Up to ten candidate articles in provider order.
3. Article title, source, publication time, URL, and snippet only when actually returned.
4. Company, round stage, amount, investors, and announcement timing only when explicitly supported by the article title or snippet; otherwise mark them unknown.
5. Duplicate or syndicated coverage grouped as one possible funding event.
6. A warning when publication recency does not establish that the financing event itself occurred inside the same window.
7. Call status, observed cost, rail, and any error.

Preserve provenance and uncertainty. Do not turn a headline into a verified deal record, invent missing fields, or claim exhaustive market coverage.

Gotchas

• No hidden defaults: both focus and freshness are required. Missing input must fail before any endpoint call.
• Publication time is not deal-close time: freshness filters news-index publication recency, not the legal close date of a financing.
• Rolling windows: pw means the past seven days and pm means the past 31 days; neither means the current calendar week or month. Apply a transparent client-side date check if the user requests an exact calendar boundary.
• Query hints are not structured filters: stage and deal-size phrases in focus improve search relevance but do not guarantee exact matches. Filter and label the returned evidence rather than fabricating database semantics.
• Press coverage is incomplete: small, private, or unannounced rounds may not appear.
• Paid failures may charge: never auto-retry. Inspect history, re-probe, and obtain fresh approval first.

Validation

• Static:

  selat skill validate ./skills/recent-funding-rounds

• Free live gate:

  SELAT_ROUTER_URL=https://router.selat.ai \
    selat skill verify ./skills/recent-funding-rounds \
    --focus "artificial intelligence" \
    --freshness "pw" \
    --live-probe

• Missing-input gate: omit focus and freshness separately and confirm each command fails before a network call.
• Paid smoke test: run one requester-approved topic only after a fresh quote, explicit approval, a spendable Gateway balance, and an armed session budget.

References

• manifest.json — machine-readable bounded payment recipe.
• references/endpoints.md — request schema, freshness semantics, live quote, scope, and free probe.
• ../../references/agent-skill-authoring-sop.md — authoring standard.
• selat-pay — https://github.com/SELAT-AI/selat-pay

"Brave" is a trademark of its respective owner and is used only for endpoint identification. This skill is not affiliated with or endorsed by Brave.