agentleFS
SELAT-AISELAT-AI/selat-skills/skills/vc-ai-infra-scout/SKILL.md

vc-ai-infra-scout

Skill2 starsChanged 42 days agoApache-2.0View on GitHub

Use this skill when the user wants VC-style deal-sourcing intelligence on AI infrastructure (inference, GPU, training, agent infra), crypto-AI / decentralized compute, robotics / embodied-AI (foundation models, humanoid, sim-to-real), or agentic payments. Runs a 9-step discovery pipeline across Tavily, Parallel (Product Hunt), Exa, SELAT-native Twitter/X, LinkedIn posts (via Tavily), and Apollo. Paid per call over mixed x402 + MPP rails through the SELAT Router. No API keys.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill when the user wants to discover early-stage companies, funding rounds, founders, and investor theses across the …

    line 17
  2. Distinctive:

    Before running, tell the user the cost: all 9 steps run every time, about $0.087 at live prices.

    line 28

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

vc-ai-infra-scout

When To Use

Use this skill when the user wants to discover early-stage companies, funding rounds, founders, and investor theses across the AI-infrastructure ecosystem. Covers four sub-theses:

  • AI Infrastructure — inference runtimes, GPU orchestration, training stacks, vector/data infra
  • Crypto-AI / Decentralized AI — on-chain AI agents, DePIN-for-compute, distributed GPU/ML
  • Robotics / Embodied AI — robotics foundation models, humanoid/embodied-AI, sim-to-real, actuation/sensor stacks
  • Agentic Payments — agent commerce, x402/stablecoin agent settlement, agentic-payment rails

Workflow

  1. Install: selat skill install vc-ai-infra-scout
  2. Pick a thesis preset (see Thesis Presets) and a target company, then run: selat skill run vc-ai-infra-scout --thesis "robotics foundation model" --twitterQuery "robotics foundation model" --fundraisingQuery "robotics startup raised seed funding" --investorQuery "robotics seed fund partner" --domain "<target-domain>". --domain is required: the run is refused without it. Before running, tell the user the cost: all 9 steps run every time, about $0.087 at live prices. Wait for a yes.
  3. The skill compiles each of 9 steps into selat-pay calls and returns a JSON response.
  4. Synthesize the 9 steps into a VC intelligence memo (tiered funding map, investor thesis extraction, geographic blindspots, strategic recommendations).

Pipeline Steps

StepSourceRailPurpose
1Tavilyx402 on BaseBroad web discovery for ${thesis} startup funding
2Parallel (Product Hunt scope)MPP on TempoProduct Hunt launches scoped to ${thesis}
3Exax402 via Circle GatewayDeep web context on startup + funding signals
4SELAT-native Twitter/X advanced_searchx402 via Circle GatewayFounder buzz, company chatter
5SELAT-native Twitter/X advanced_searchx402 via Circle GatewayRecent pre-seed/seed fundraising news
6Tavily (LinkedIn scope)x402 on BaseLinkedIn post search for funding announcements
7SELAT-native Twitter/X advanced_searchx402 via Circle GatewayInvestor/fund-partner thesis tweets
8Apollo people-searchMPP on TempoFounder shortlist by keyword + title
9Apollo org-enrichmentMPP on TempoDeep-dive enrichment on the target domain

Rail labels name each merchant's rail, not the chain you pay from. The SELAT Router settles every step from your funded Circle Gateway balance.

Thesis Presets

The manifest has no preset mechanism, so each preset is a set of flags to copy. Always add --domain.

Track--thesis--twitterQuery--fundraisingQuery--investorQuery
AI infrastructureAI infrastructureAI infra founderAI infra startup raised seedAI infra VC partner thesis
Robotics / embodied AIrobotics foundation modelrobotics foundation modelrobotics startup raised seed fundingrobotics seed fund partner
Crypto-AI / DePINdecentralized AI computedecentralized AI compute founderdecentralized AI startup raised seedcrypto AI VC partner thesis
Agentic paymentsagentic payment railsagentic payments founderagentic payment startup raiseagentic payments VC partner thesis

For robotics, avoid the phrase "embodied AI" in Twitter queries (see Gotchas).

Inputs And Outputs

ParamRequiredDefaultDescription
thesisnoAI infrastructureCore search term. Try: robotics foundation model, humanoid robotics, decentralized AI compute, agentic payment rails
twitterQuerynoAI infra founderTwitter/X search for founder buzz. Override for your thesis: e.g. robotics foundation model
fundraisingQuerynostartup raised seed pre-seed funding roundFundraising news query. Override: robotics startup raised seed funding
investorQuerynoseed fund partner thesisInvestor/partner chatter. Override: robotics seed fund partner
domainyesnoneDomain of the one company to enrich in step 9 (Apollo org-enrichment). No default, so a placeholder can never silently enrich an unrelated company.

Output: JSON with ok, skill, user_summary, and steps[] array containing per-step API responses.

Gotchas

  • Intent fidelity: Robotics foundation-model queries surface less on general web search and Product Hunt than infra queries do. Steps 1–3 may miss early-stage robotics companies that are stealth or academic. Supplement with a direct selat-pay call to SELAT-native Twitter/X advanced_search when needed.
  • Timeout leakage: A selat skill run vc-ai-infra-scout that times out can leave in-flight micropayments settling against the session budget. Use selat-pay directly for single follow-up calls.
  • All 9 steps always run: selat skill run has no step selection, so every run pays for all 9 calls. Optional params you omit fall back to their AI-infra defaults; pass all four query params for a different thesis.
  • Domain is required: Step 9 enriches exactly the --domain you pass, and the run is refused without one (v1.2.0 removed the old modal.com default, which enriched Modal on every run regardless of thesis). Every run pays for the step 9 call, so pass a company you actually want enriched: one you already track, or a lead from an earlier free selat search or web search. The pipeline cannot pick it for you, because all 9 steps run in one pass.
  • Embodied AI null on Twitter: The phrase "embodied AI" frequently returns zero tweets. Use physical AI, foundation model for robots, or company-specific hashtags instead.

Validation

--chain base in the probe command below is only the flag selat-pay requires today — a probe reads a free, chain-independent quote and never settles. A real paid run resolves the settlement chain from your funded Circle Gateway balance, not the manifest.

  • Static: selat skill validate ./skills/vc-ai-infra-scout
  • Live gate (free, probe-only): selat skill verify ./skills/vc-ai-infra-scout --live-probe --domain figure.ai
  • Single-step probe (no pay): selat-pay POST "https://apollo.mpp.paywithlocus.com/apollo/org-enrichment" --body '{"domain":"figure.ai"}' --chain base --probe-only --live-probe
  • Full run (paid): selat skill run vc-ai-infra-scout --thesis "<your thesis>" --domain "<target-domain>" --json
  • A successful run prints status=200 for each step.

Cost Estimate

StepEndpointLive quote (probed 2026-10-04)Step cap
1Tavily$0.0105$0.02
2Parallel$0.0105$0.05
3Exa$0.007$0.05
4–5–7SELAT-native Twitter/X ×3$0.001 × 3 = $0.003$0.05 each
6Tavily (LinkedIn)$0.0105$0.02
8Apollo people-search$0.00525$0.05
9Apollo org-enrichment$0.0399$0.06
Total~$0.087 (two paid runs on 2026-10-02 cost $0.0867 each)

Each step is capped by its own maxAmount. The top-level maxAmount ($0.06) is only a fallback for a step with no cap of its own; it is not a total-run cap. The worst case is the sum of the step caps ($0.40).

References

  • manifest.json — the machine-readable payment recipe this skill runs.
  • references/endpoints.md — catalogue endpoints called by this skill.
  • references/agent-skill-authoring-sop.md — authoring standard.
  • selat-pay — https://github.com/SELAT-AI/selat-pay

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.