SuperAPI AI Directory

web3-opportunities

General & other

View repository

Updated 10-22 16:15

About this skill

web3-opportunities

Answers "where can a Web3 builder/founder get funded or launch?" from a curated, bundled roster of accelerators, incubators, grants, hackathons, bounties, retroactive funding, and fellowships. The roster's stable facts (what a program is, who it serves, which chains and regions, dilutive vs non-dilutive) ship offline. Time-sensitive details (deadlines, open cohorts, current prize pools) are baseline snapshots that the agent verifies live before the user acts.

Tool-name note for portability: this skill body uses python3 for the shell/terminal, Read for file reads, and WebSearch/WebFetch for live web lookups. Map these to your runtime's equivalents if they are named differently.

When to Use

Trigger this skill when the user wants opportunities, not regulation:

  • "Where can I get a Web3/crypto grant?" / "non-dilutive funding for [chain] project"
  • "Crypto accelerators / incubators for my stage" / "which accelerator should I apply to?"
  • "Hackathons I can enter" / "ETHGlobal / Colosseum / Solana hackathon schedule"
  • "Ecosystem grants for Arbitrum / Optimism / Base / Polkadot / Cosmos / NEAR / Sui / Filecoin"
  • "RetroPGF / retroactive funding" / "bounties for builders" / "Web3 fellowships"
  • "Funding options for a Southeast Asia / Vietnam crypto founder"

This skill is about funding and launch ramps. For Vietnam crypto law, tax, and licensing, use the vietnam-crypto-radar skill instead.

Core method: catalog first, enrich second

  1. Query the bundled roster with the user's facets:
    python3 "$SKILL_DIR/scripts/query_opportunities.py" --type grant --dilution non-dilutive --region sea
    
    The script returns matching entries plus data_as_of and time_sensitive_fields.
  2. Trust the stable fields as-is: type, stage, dilution, chains, regions, url, name. These are evergreen.
  3. Treat the time-sensitive fields as baseline snapshots: status, cadence, typical_check_or_prize, and any deadline. They reflect what was true on the entry's last_verified date, not necessarily today.
  4. Enrich when the user needs to act (apply, hit a deadline, confirm a prize). Use WebSearch/WebFetch against the entry's url (Tier-1) and references/sources.md.
  5. Label every time-sensitive fact: mark it [bundled baseline · as of <last_verified>] or [live-verified · <today>]. Never present a baseline deadline or "open" status as current without checking.

Usage

# Full roster (default when no filters are given)
python3 "$SKILL_DIR/scripts/query_opportunities.py" --all

# Non-dilutive grants for an idea-stage founder in Southeast Asia
python3 "$SKILL_DIR/scripts/query_opportunities.py" --type grant --dilution non-dilutive --stage idea --region sea

# All Solana hackathons
python3 "$SKILL_DIR/scripts/query_opportunities.py" --type hackathon --chain solana

# SEA / Vietnam highlight surface (curated relevance flag)
python3 "$SKILL_DIR/scripts/query_opportunities.py" --sea

# OR within a facet, AND across facets
python3 "$SKILL_DIR/scripts/query_opportunities.py" --type grant,hackathon --chain ethereum,l2 --region global

# Free-text search over name / id / notes
python3 "$SKILL_DIR/scripts/query_opportunities.py" --search optimism

Filtering is AND across facets, OR within a facet. List facets (--type, --stage, --chain, --region, --status) accept repeated flags or comma-separated values. --dilution is exact-match: non-dilutive does NOT auto-include mixed — ask for mixed explicitly when relevant. If $SKILL_DIR is not set by the runtime, replace it with the absolute path to this skill directory.

Verification discipline

  • The roster is the source of truth for a program's EXISTENCE and STABLE facets. Do not enumerate programs from memory — run the script.
  • Time-sensitive fields carry a per-entry last_verified anchor; the set carries data_as_of. State the baseline date when you quote them unverified.
  • A time-sensitive fact counts as "live-verified" only if confirmed on the program's official page (Tier-1) or two independent aggregators (Tier-2) — see references/sources.md.
  • When live data contradicts the baseline, live wins; note the drift so it can be fixed in the data file later.
  • This is informational, not financial or legal advice.

Output / presentation template

Present results grouped by type, with a separate SEA/Vietnam highlight when relevant:

# Web3 Opportunities — <facet summary>
_Bundled baseline as of <data_as_of>. Verify time-sensitive fields (*) live before applying._

## Grants (non-dilutive)
- **<name>** — <chains> · <regions> · <stage> — <check/prize>* — <status>* — <url>
...

## Accelerators (dilutive)
- ...

## Hackathons / Bounties / Retroactive funding / Fellowships / Education
- ...

## SEA / Vietnam highlight
- <entries where sea_relevant = true; includes VN student/education programs
  (Corelia Academy, UniHackfest, VBI Academy) and community ramps (Superteam Vietnam)>

## Next steps
- The 2-3 best fits to verify live now (deadline / open cohort), and which source to check.
(* = time-sensitive baseline snapshot)

Critical Rules

  • ALWAYS run query_opportunities.py before listing opportunities. Never invent programs.
  • NEVER present status, cadence, typical_check_or_prize, or a deadline as current without either labeling it baseline-as-of-last_verified or live-verifying it.
  • When the query returns zero matches, say so plainly and offer the live-lookup path via references/sources.md rather than fabricating entries or loosening facts silently.
  • When the user is regionally focused (especially SEA / Vietnam), surface the sea_relevant entries explicitly.
  • If the roster is missing a program the user names, find it live via the Tier-2 aggregators in references/sources.md and offer to add it to the data file.

Reference files

  • data/web3_opportunities.json — the bundled roster (stable fields + time-sensitive baselines + facet enums).
  • references/sources.md — tiered registry of official program pages and aggregators for refreshing time-sensitive fields and discovering programs not yet in the roster.

Prerequisites

  • Python 3.10+. No API key required for the bundled catalog (fully offline).
  • The live-enrichment layer needs web access (WebSearch/WebFetch); it is optional — the catalog answers structural questions on its own.