agentleFS
JerryLiu369JerryLiu369/awesome-quant-skills/skills/crypto-web3/defillama-data/SKILL.md

defillama-data

Skill3 starsChanged 33 days agoMITView on GitHub

Many protocols have multiple versions. The database tracks both: - Parent protocol (e.g., aave) — used in get_protocol_metrics, aggregates all versions - Sub-protocol (e.g., aave-v3) — individual versions, used in sub_protocol views resolve_entity returns both slug (sub-protocol) and parent_protocol. Use parent_protocol for get_protocol_metrics. Use slug when you need a specific version. Example: resolve_entity({name: "compound"}) returns: - slug: compound-v2, parent_protocol: compound - slug: compound-v3, parent_protocol: compound → Use protocol: "compound" in get_protocol_metrics for the aggregate. All entity params accept single values…

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    All entity params accept single values or arrays for comparison:

    line 67
  2. Very specific:

    All tools accept period (NOT time_range):

    line 83
  3. Very specific:

    FDV: Never sort by FDV -- unreliable total_supply data produces junk rankings.

    line 139

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
<!-- Imported from: https://github.com/DefiLlama/defillama-skills Source Path: defi-data/SKILL.md Curated for awesome-quant-skills -->

name: defi-data description: Core reference for DefiLlama MCP tools. Maps DeFi questions to the correct tool call with proper parameters. Covers entity conventions, metric interpretation, stock vs flow distinctions, percentage formatting, and error recovery. Use whenever querying DeFi data — protocol TVL, token prices, chain metrics, fees, revenue, yields, stablecoins, bridges, ETFs, hacks, raises, treasuries, or institutional holdings.

DefiLlama MCP Data Reference

Tool Quick Reference

QuestionToolKey Params
Total DeFi TVL, global volumesdefillama:get_market_totalsmetrics (tvl_base, volume_dexs_1d, etc.)
Protocol TVL, fees, revenuedefillama:get_protocol_metricsprotocol, metrics
Chain TVL, gas fees, REVdefillama:get_chain_metricschain, metrics
Token price, mcap, ATHdefillama:get_token_pricestoken
Category rankingsdefillama:get_category_metricscategory
Pool APY, lending ratesdefillama:get_yield_poolstoken, chain, category
Stablecoin supplydefillama:get_stablecoin_supplystablecoin, chain
ETF inflows/outflowsdefillama:get_etf_flowstoken (bitcoin/ethereum)
Bridge volumedefillama:get_bridge_flowsbridge, chain
CEX volumedefillama:get_cex_volumescex
Open interestdefillama:get_open_interestprotocol, chain
Hacks/exploitsdefillama:get_eventsevent_type: "hacks"
Fundraisingdefillama:get_eventsevent_type: "raises"
Protocol eventsdefillama:get_eventsevent_type: "protocol_events"
Protocol treasurydefillama:get_treasurytreasury
Institutional holdingsdefillama:get_dat_holdingsinstitution, token
Token unlocksdefillama:get_token_unlockstoken, query_type (chart, window, day, window_all, ratio, market_window)
Token TVL in protocolsdefillama:get_token_tvltoken, protocol
Oracle coveragedefillama:get_oracle_metricsoracle
User activitydefillama:get_user_activityprotocol, chain
Income statementdefillama:get_income_statementprotocol
Protocol metadata, URLs, auditdefillama:get_protocol_infoprotocol, category, tag, has_token, chain
Chain metadata, type, L2 parentdefillama:get_chain_infochain, chain_type
List valid categoriesdefillama:list_categoriestype (protocol, token, chain)
Unknown entity slugdefillama:resolve_entityname

Entity Conventions

  • Protocols: lowercase-hyphenated slugs (aave, uniswap, compound-v3)
  • Chains: lowercase (ethereum, solana, arbitrum)
  • Tokens: coingecko: prefix (coingecko:ethereum, coingecko:bitcoin, coingecko:usd-coin)
  • CEX: slug includes suffix (binance-cex, coinbase-cex, htx, gate-io)
  • Categories: Title case (Lending, DEX, Derivatives)

Protocol vs Sub-Protocol

Many protocols have multiple versions. The database tracks both:

  • Parent protocol (e.g., aave) — used in get_protocol_metrics, aggregates all versions
  • Sub-protocol (e.g., aave-v3) — individual versions, used in sub_protocol views

resolve_entity returns both slug (sub-protocol) and parent_protocol. Use parent_protocol for get_protocol_metrics. Use slug when you need a specific version.

Example: resolve_entity({name: "compound"}) returns:

  • slug: compound-v2, parent_protocol: compound
  • slug: compound-v3, parent_protocol: compound → Use protocol: "compound" in get_protocol_metrics for the aggregate.

Array Support

All entity params accept single values or arrays for comparison:

  • protocol: "aave" or protocol: ["aave", "lido", "compound"]
  • chain: "ethereum" or chain: ["ethereum", "solana"]
  • token: "coingecko:ethereum" or token: ["coingecko:ethereum", "coingecko:bitcoin"]

Entity Resolution

Try tools directly for well-known entities — works 80%+ of the time. Use defillama:resolve_entity only when:

  • Unsure of exact slug
  • Tool returns 0 rows
  • Need to disambiguate (e.g., "compound" → compound-v2, compound-v3)
  • Need to find the parent_protocol slug

Period Parameter

All tools accept period (NOT time_range):

  • current (default) -- latest snapshot
  • 7d, 30d, 90d, 180d, 365d -- rolling windows

For custom date ranges, use start_date and end_date (YYYY-MM-DD format) instead of period. If start_date is set without end_date, it defaults to today.

Critical Conventions

Stock vs Flow

Different aggregation rules -- getting this wrong produces nonsensical results:

  • Stock (point-in-time snapshots): tvl, price, mcap -- NEVER sum across dates
  • Flow (cumulative over period): fees_1d, revenue_1d, volume -- CAN sum across dates

Percentage Formatting

Two different conventions in the data:

  • *_pct_change columns = decimals (e.g., -0.13 = -13%, 0.35 = +35%). Multiply by 100 for display.
  • apy columns = already percentages (e.g., 2.32 = 2.32%). Do NOT multiply by 100.

TVL Components

Components are mutually exclusive -- never double-add:

  • tvl_base = default TVL (excludes liquid staking and double-counted)
  • tvl_ls_only = liquid staking portion not in tvl_base
  • tvl_dc_only = double-counted portion not in tvl_base
  • tvl_ls_and_dc = both LS and DC
  • Total with LS+DC = tvl_base + tvl_ls_only + tvl_dc_only + tvl_ls_and_dc

Fee and Revenue

  • fees = gross payments by users
  • revenue = protocol's cut (fees - payments to LPs/stakers)
  • holder_revenue = buybacks + dividends to token holders
  • Chain-level: chain_fees (gas), chain_revenue, chain_mev, app_fees (protocol fees on chain), app_revenue

Metric Name Patterns

Protocol metrics follow these patterns:

  • TVL: tvl_base, tvl_staking, tvl_borrowed, tvl_pool2, tvl_treasury, tvl_vesting
  • Flows (with suffixes _1d/_7d/_30d/_90d/_180d/_365d/_ytd/_alltime/_annualised): fees, revenue, holder_revenue, incentive
  • Volume (same suffixes): volume_dexs, volume_aggregators, volume_derivatives, volume_aggr_derivatives, volume_options
  • Market: price, mcap, fdv, fdv_outstanding, token_volume, ps_ratio, pf_ratio
  • pct_change variants only available on _current views

Chain metrics use chain_fees, chain_revenue, chain_mev, app_fees, app_revenue instead of fees, revenue.

Daily views only have the _1d suffix for flows/volumes.

Important Gotchas

  • Category column: get_protocol_metrics has NO category column. Use get_category_metrics for category-level data.
  • FDV: Never sort by FDV -- unreliable total_supply data produces junk rankings. Sort by mcap instead.
  • Valuation: ps_ratio = mcap / annualised revenue. pf_ratio = mcap / annualised fees. Lower = potentially undervalued.
  • Metric validation: Protocol/chain/category metrics are validated against actual DB columns. Invalid metrics return a suggestion for the closest match.
  • NULL values: NULL means data unavailable, NOT $0. Do not report NULL as zero.
  • Params accept string or array: protocol: "aave" or protocol: ["aave", "lido"]. Param names are singular (protocol not protocols).
  • sort_by format: String like "tvl_base desc" or "fees_1d desc".

Examples

Example 1: User: "What is Aave's TVL?" Tool call: defillama:get_protocol_metrics(protocol: "aave", metrics: ["tvl_base"])

Example 2: User: "Top 5 chains by fees" Tool call: defillama:get_chain_metrics(sort_by: "chain_fees_1d desc", limit: 5)

Example 3: User: "ETH price history over 90 days" Tool call: defillama:get_token_prices(token: "coingecko:ethereum", period: "90d")

Example 4: User: "Compare ETH price and Ethereum chain TVL" Tool calls in parallel:

  • defillama:get_token_prices(token: "coingecko:ethereum")
  • defillama:get_chain_metrics(chain: "ethereum")

Example 5: User: "Biggest DeFi hacks this year" Tool call: defillama:get_events(event_type: "hacks", sort_by: "amount desc", limit: 10, period: "365d")

Example 6: User: "Aave TVL from Jan to Mar 2025" Tool call: defillama:get_protocol_metrics(protocol: "aave", metrics: ["tvl_base"], start_date: "2025-01-01", end_date: "2025-03-31")

Error Recovery

  1. 0 rows returned -> call defillama:resolve_entity(name: "...") -> retry with the correct slug
  2. Wrong data -> verify you're using the right tool (protocol vs chain vs category)
  3. Ambiguous entity -> call resolve_entity to see all matching slugs, then pick the right one

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.