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
- SELAT-AI/selat-openclaw-skillstheir own version, first seen Jul 26, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- SELAT-AI/selat-openclaw-skills★ 2same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository19 other files it gives its agents.
- account-intelSkill
- comprehensive-enrichmentSkill
- email-campaignSkill
- enrich-waterfallSkill
- financial-intelSkill
- find-twitter-influencersSkill
- gtm-enrichment-deepSkill
- gtm-enrichment-smartSkill
- lead-enrichmentSkill
- perplexity-searchSkill
- person-lookupSkill
- recent-funding-roundsSkill
- sales-prospectingSkill
- scrapecreatorsSkill
- self-evolving-agentSkill
- social-intelSkill
- stock-direction-signalsSkill
- twitter-researchSkill
- wallet-desk-briefSkill
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
- Distinctive: line 17
Use this skill when the user wants to discover early-stage companies, funding rounds, founders, and investor theses across the …
- Distinctive: line 28
Before running, tell the user the cost: all 9 steps run every time, about $0.087 at live prices.
≈ “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.
Files
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
- Install:
selat skill install vc-ai-infra-scout - 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>".--domainis 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. - The skill compiles each of 9 steps into
selat-paycalls and returns a JSON response. - Synthesize the 9 steps into a VC intelligence memo (tiered funding map, investor thesis extraction, geographic blindspots, strategic recommendations).
Pipeline Steps
| Step | Source | Rail | Purpose |
| 1 | Tavily | x402 on Base | Broad web discovery for ${thesis} startup funding |
| 2 | Parallel (Product Hunt scope) | MPP on Tempo | Product Hunt launches scoped to ${thesis} |
| 3 | Exa | x402 via Circle Gateway | Deep web context on startup + funding signals |
| 4 | SELAT-native Twitter/X advanced_search | x402 via Circle Gateway | Founder buzz, company chatter |
| 5 | SELAT-native Twitter/X advanced_search | x402 via Circle Gateway | Recent pre-seed/seed fundraising news |
| 6 | Tavily (LinkedIn scope) | x402 on Base | LinkedIn post search for funding announcements |
| 7 | SELAT-native Twitter/X advanced_search | x402 via Circle Gateway | Investor/fund-partner thesis tweets |
| 8 | Apollo people-search | MPP on Tempo | Founder shortlist by keyword + title |
| 9 | Apollo org-enrichment | MPP on Tempo | Deep-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 infrastructure | AI infrastructure | AI infra founder | AI infra startup raised seed | AI infra VC partner thesis |
| Robotics / embodied AI | robotics foundation model | robotics foundation model | robotics startup raised seed funding | robotics seed fund partner |
| Crypto-AI / DePIN | decentralized AI compute | decentralized AI compute founder | decentralized AI startup raised seed | crypto AI VC partner thesis |
| Agentic payments | agentic payment rails | agentic payments founder | agentic payment startup raise | agentic payments VC partner thesis |
For robotics, avoid the phrase "embodied AI" in Twitter queries (see Gotchas).
Inputs And Outputs
| Param | Required | Default | Description |
thesis | no | AI infrastructure | Core search term. Try: robotics foundation model, humanoid robotics, decentralized AI compute, agentic payment rails |
twitterQuery | no | AI infra founder | Twitter/X search for founder buzz. Override for your thesis: e.g. robotics foundation model |
fundraisingQuery | no | startup raised seed pre-seed funding round | Fundraising news query. Override: robotics startup raised seed funding |
investorQuery | no | seed fund partner thesis | Investor/partner chatter. Override: robotics seed fund partner |
domain | yes | none | Domain 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-paycall to SELAT-native Twitter/X advanced_search when needed. - Timeout leakage: A
selat skill run vc-ai-infra-scoutthat times out can leave in-flight micropayments settling against the session budget. Useselat-paydirectly for single follow-up calls. - All 9 steps always run:
selat skill runhas 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
--domainyou pass, and the run is refused without one (v1.2.0 removed the oldmodal.comdefault, 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 freeselat searchor 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 basein the probe command below is only the flagselat-payrequires 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=200for each step.
Cost Estimate
| Step | Endpoint | Live quote (probed 2026-10-04) | Step cap |
| 1 | Tavily | $0.0105 | $0.02 |
| 2 | Parallel | $0.0105 | $0.05 |
| 3 | Exa | $0.007 | $0.05 |
| 4–5–7 | SELAT-native Twitter/X ×3 | $0.001 × 3 = $0.003 | $0.05 each |
| 6 | Tavily (LinkedIn) | $0.0105 | $0.02 |
| 8 | Apollo people-search | $0.00525 | $0.05 |
| 9 | Apollo 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
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
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.

