AlgoVault Skills
May 18, 2026 · View on GitHub
20 Anthropic Agent Skills over the AlgoVault MCP server.
Each Skill is a single-prompt wrapper over 1–3 calls to AlgoVault's signal-interpretation API
(get_market_regime, get_trade_signal, scan_funding_arb). Install once, invoke from any
Claude Code or Cowork session, and we surface the composite verdict — your agent decides execution.
AlgoVault provides the thesis; your agent decides execution.
Quick start
# In Claude Code
claude plugin install AlgoVaultLabs/algovault-skills
# In Cowork — discover via the marketplace, click "install"
Live track record (PFE Win Rate + cumulative call count, updated continuously): https://algovault.com/track-record
Integrations
End-to-end tutorials pairing AlgoVault's analytics with each major exchange's agent execution kit. AlgoVault returns the verdict; the agent and its risk policy decide what (if anything) to execute. All demos run on testnet/demo mode — zero real-money risk.
Quick install + run any demo:
claude plugin install AlgoVaultLabs/algovault-skills
git clone https://github.com/AlgoVaultLabs/algovault-skills && cd algovault-skills && npm install
BINANCE_TESTNET=true node examples/binance/demo.mjs # or OKX_DEMO=true / BYBIT_TESTNET=true / BITGET_DEMO=true
Framework integrations
Drop-in tutorials for the major Python agent frameworks. Each tutorial pairs
AlgoVault MCP with the framework's canonical MCP-adapter library — copy-pasteable
demo code, no AlgoVault SDK required. python examples/<framework>/demo.py BTC 4h
returns a real BUY/SELL/HOLD verdict from api.algovault.com/mcp.
The 20 Skills
| # | Slug | Name | Difficulty | Tools |
|---|---|---|---|---|
| 01 | quick-btc-check | Quick BTC Check | Beginner | get_trade_signal |
| 02 | portfolio-scanner | Portfolio Scanner | Intermediate | get_trade_signal |
| 03 | regime-aware-trading | Regime-Aware Trading | Intermediate | get_market_regime, get_trade_signal |
| 04 | funding-arb-monitor | Funding Arb Monitor | Intermediate | scan_funding_arb |
| 05 | full-3-tool-pipeline | Full 3-Tool Pipeline | Advanced | get_market_regime, get_trade_signal, scan_funding_arb |
| 06 | multi-timeframe-confirmation | Multi-Timeframe Confirmation | Advanced | get_trade_signal |
| 07 | tradfi-rotation | TradFi Rotation | Advanced | get_market_regime, get_trade_signal |
| 08 | risk-gated-entry | Risk-Gated Entry | Advanced | get_market_regime, get_trade_signal |
| 09 | funding-sentiment-dashboard | Funding Sentiment Dashboard | Advanced | get_market_regime |
| 10 | contrarian-meme-scanner | Contrarian Meme Scanner | Advanced | get_market_regime, get_trade_signal |
| 11 | divergence-detector | Divergence Detector | Advanced | get_market_regime, get_trade_signal |
| 12 | hourly-digest-bot | Hourly Digest Bot | Advanced | get_trade_signal, get_market_regime |
| 13 | hedging-advisor | Hedging Advisor | Advanced | get_market_regime, get_trade_signal, scan_funding_arb |
| 14 | volatility-breakout-watch | Volatility Breakout Watch | Advanced | get_market_regime, get_trade_signal |
| 15 | cross-asset-correlation | Cross-Asset Correlation | Advanced | get_trade_signal |
| 16 | funding-cash-and-carry | Funding Cash-and-Carry | Advanced | scan_funding_arb, get_trade_signal |
| 17 | weekend-vs-weekday-patterns | Weekend vs Weekday Patterns | Research | get_trade_signal, get_market_regime |
| 18 | agent-portfolio-rebalance | Agent Portfolio Rebalance | Advanced | get_market_regime |
| 19 | smart-dca-bot | Smart DCA Bot | Advanced | get_trade_signal |
| 20 | multi-agent-war-room | Multi-Agent War Room | Expert | get_market_regime, get_trade_signal, scan_funding_arb |
How attribution works
Every Skill invocation sends X-AlgoVault-Skill-Slug: <slug> to the MCP server. We aggregate
these in an admin-only counter inside /dashboard on api.algovault.com (per-Skill funnel
data is competitive intel — moved internal 2026-04-24). The public moat layer remains the
aggregate track record at https://algovault.com/track-record (Merkle-verified PFE WR +
signal/batch counts).
Repo layout
algovault-skills/
├── .claude-plugin/
│ ├── plugin.json # Plugin manifest (mcpServers, version, etc.)
│ └── marketplace.json # Marketplace entry
├── skills/
│ ├── _template/ # SKILL.md.tmpl — generator template
│ ├── manifest.json # Single source of truth: 20 Skill definitions
│ ├── manifest.schema.json # JSON Schema v7 for manifest validation
│ └── <slug>/SKILL.md # GENERATED — never hand-edit
├── scripts/
│ ├── gen-skill.mjs # Reads manifest, emits SKILL.md per entry
│ └── skills_preflight.sh # Pre-build environment check
├── tests/smoke/ # Live invocation tests against api.algovault.com
└── docs/ # HOW_TO_ADD.md + integrations/{binance,okx,bybit,bitget}.md + SUBMIT_<EXCHANGE>.md templates
Development
npm run preflight # Verify env + repo + MCP reachable
npm run gen # Regenerate all SKILL.md files from manifest
npm test # Live smoke test against api.algovault.com
npm run test:unit # Offline unit tests (helper + template-consistency canary)
Submission generator
scripts/generate_skill_submission.mjs reads a docs/SUBMIT_<EXCHANGE>.md
template, fetches live track-record data from the AlgoVault API, and emits
ready-to-paste submission artifacts (SKILL.md for Skills-Hub PRs,
PR_TITLE.txt, PR_BODY.md, SUMMARY.md for every exchange):
node scripts/generate_skill_submission.mjs --exchange BINANCE
node scripts/generate_skill_submission.mjs --exchange BYBIT --output-dir /tmp/foo/
Templates use <PLACEHOLDER> tokens (<PFE_WR>, <TOTAL_CALLS>,
<BATCH_COUNT>, <VERSION>, etc.) so submission text never carries stale
hardcoded numbers. CI canary tests/unit/submit-template-consistency.test.mjs
fails the build if any template re-introduces a literal — keeps the templating
discipline locked in.
Contributing
Skills are added via skills/manifest.json only — never hand-edit generated SKILL.md files.
Add a manifest entry, run npm run gen, commit both. PRs welcome.
Also built by AlgoVault Labs
@algovaultofficialbot— free public Telegram bot. Regime alerts + AlgoVault trade calls (BUY/SELL) pushed to your watchlist. Same composite-verdict signal stream these skills wrap. Source: github.com/AlgoVaultLabs/algovault-bot.
Built by AlgoVault Labs
Composable signal-interpretation tools for AI trading agents — MIT licensed.