Skills Catalog

Recommends a 3-validator NEAR delegation portfolio scored on commission, uptime, slashing history, and decentralization (avoid concentration in a single nodeprovider). Different from v1 governance — focused on yield, not voting.

Sweeps stablecoin supply rates across NEAR-side + Aurora-side lending markets; surfaces the best risk-adjusted venue for a given size. Optionally fires the move via 1-click bridge.

Auto-claims and re-stakes LP rewards on a schedule, gas-aware (won't fire if claim cost > N% of rewards). Logs APY uplift vs. unclaimed baseline.

For a given size, finds the best liquid-staking entry route (direct mint vs. secondary buy) factoring discount-to-NAV and slippage. Stops "I bought stNEAR at a 1.5% premium and didn't notice."

Logs every staking-reward receipt with USD basis at receipt time across all your validators / LST positions / lending markets. The thing your accountant will actually thank you for.

Rebalances across LSTs (stNEAR / LiNEAR / meta-pool) and direct stakes to target weights. Different from v1's NewsCoin rebalancer — focuses on yield-bearing assets, not story tokens.

For an LP position, recommends an offsetting short (perp or borrow-and-sell) sized to neutralize first-order IL exposure. Optional auto-execute with size cap.

Predicts when a yield farm's emissions decay below your hurdle rate (commit a number); auto-exits when triggered. Specific to ramp-down farms (most yield comes early; many users overstay).

Claims staking rewards across multiple validators on one cron. Threshold-gated to skip dust. Different from v1's NewsCoin creator-fee claimer — for native staking, not token economics.

Single-card view of every unbonding/cooldown timer across your chains and protocols (NEAR validator unstake, Aave cooldown, etc.). Pings you on the day funds become withdrawable so they don't sit idle.

For a chosen asset, finds chains where the same asset earns higher yield, factoring 1-click bridge fees + lockup risk. Surfaces actual after-cost APY.

Tracks vesting unlocks (yours + watched tokens) and projects supply-impact on price. Differs from generic unlock calendars by quantifying expected sell pressure from prior cohort behavior.

Visualizes the cross-protocol yield curve for an asset (e.g. USDC: 4% on Aave, 7% on Burrow, 11% on a farm) and flags anomalies (curve inversions, fat-tail rates) that usually mean risk.

Allocates a stablecoin bag across yield venues by risk-adjusted return (target volatility cap + concentration limits). Rebalances quarterly. The "park your stables" service.

Maps your restaked positions across protocols (EigenLayer/Karak/Symbiotic via Aurora) and the AVS/services they secure, with combined slashing exposure. Surfaces "you're stacking 3x slashing risk" before it bites.

Sweeps N items at the floor of a chosen collection within budget, with optional rarity floor (don't grab a 1/1 by accident). Protects against floor manipulation by capping fill velocity.

Scores an NFT 0–100 on rarity within its collection (trait-rarity composite) and ranks vs. its current ask price. Catches under-priced rares and over-priced commons.

Alerts on mint windows for watched collections via on-chain detection (new collection contract from a creator you watch) — earlier than calendar-based alerts.

Daily MTM your NFT bag using floor + last-sale + recent-comp weighted estimate. Different from generic floor multiplication — handles rare items separately.

Tracks creator royalties earned across marketplaces (where enforced), broken down by collection, marketplace, and time window. Helps creators benchmark which marketplaces actually pay.

Recommends list price + duration based on collection liquidity (depth at floor, average days-to-sale, recent comp velocity). Stops you parking at the floor for 90 days when 2% above sells in 3.

Places tiered bids on a collection at floor / floor-5% / floor-10% with auto-renewal. Different from sniping — passive accumulation strategy for collections you'll hold.

Finds NFTs whose ask price is below their trait-floor (the floor for items sharing their rare trait). Often a lazy lister or a trait-floor mover from a recent sale.

Cross-references your NFT holdings against ongoing holder-airdrops you may be eligible for (NEAR + bridged EVM). One-tap claim where supported.

Finds NFT-collateralized loan offers against pieces in your bag (LTV, APR, duration ranking). Doesn't take the loan — just lays out options when you want liquidity without selling.

Tracks P2P NFT derivative exposures: rentals, fractional, options. Stops the "I forgot I rented out my NFT" surprise on day 30.

Flags wash-traded volume on a collection before you buy in (round-trip trades between a small set of wallets, flat-price trades, suspicious gas patterns).

Daily DM/TG digest of upcoming mints across your watched chains, ranked by team-prior + community signals. Different from `nft-mint-drop-alert` — calendar-based + curated, not real-time on-chain.

Aggregates yields across NFT-staking programs you're eligible for (or could be, if you bought-in). Surfaces the "stake to earn $X token" landscape per chain.

Single dashboard of pending votes across every DAO you hold tokens in (NEAR + Aurora + Snapshot-based EVM). One screen, every deadline.

Daily delta on every watched DAO treasury (inflow/outflow/composition). Catches "the treasury moved $5M to a new wallet" before it hits the news.

Finds a delegate active across multiple DAOs whose vote pattern aligns with your stated values. Different from v1's AZUKA-only delegate finder — cross-DAO scope.

Tracks Snapshot proposals you're eligible to vote on (token-gated by holding). One-click vote where Snapshot supports it.

Auto-casts your vote per a stated policy on standardized proposal types (e.g. "always for sub-$50k grants under approved budget", "always against new-token issuance"). Reviewed by you weekly.

Tracks outstanding grant payments across DAOs you've contracted with. The "I delivered the work, where's the wire" service.

Lists every cross-DAO proposal short of quorum within 24h of close. Different from v1's AZUKA-only quorum-watcher.

Aggregates working-group calls and forum discussions across DAOs into your calendar. Discord/Telegram/forum events into one ICS.

Lists open bounties across DAOs matching your skills. Cross-DAO version of v1's mission-finder.

Claims retro/airdrops/rewards across DAOs on schedule. Threshold-gated to skip dust. Saves the gas-vs-value math.

Executes batched cross-DAO operations (vote on N proposals + claim on M + delegate K) in minimum-tx form, factoring per-chain gas.

Flags when two DAOs you're in voted opposite ways on related questions (e.g. "DAO A funded project X, DAO B funded its competitor Y"). Useful for accountability or just understanding the field.

Daily summary of forum threads in DAOs you watch (Discourse/Commonwealth). Beats refreshing 8 forums to see what's brewing pre-proposal.

Tracks your effective voting power per DAO over time (token balance + delegations received + lock multipliers). Stops the "wait, I had more votes last month" surprise.

Calculates realized gains FIFO across your full wallet history including DeFi positions and bridges. Different from v1's NewsCoin-only tax export — full multi-chain scope.

Explains crypto tax rules for the user's stated jurisdiction (US/UK/EU/AU/SG) with citations to actual regs. Updated when guidance changes.

Classifies staking income vs capital gain per jurisdiction. Critical for jurisdictions where staking is income at receipt vs cap-gain at sale (most are the former).

Finds unrealized loss positions ripe for tax-loss-harvest before year-end, with wash-sale advisor (e.g. don't rebuy same token within 30d in US).

Computes USD basis at receipt for every airdrop you've received, using closest-trade pricing. The thing every tax tool gets wrong by 10–30%.

Pre-fills US Form 8949 / UK SA108 / EU equivalents from on-chain data. Doesn't file — produces the populated PDF.

Consolidates trades across N wallets into one ledger, marking internal transfers (not taxable). Solves the "I have 4 wallets and Koinly thinks every transfer is a sale" problem.

Auto-flags transfers between your own wallets across chains so they're excluded from gains. Differs from `accounting-multi-wallet-consolidate` — runs continuously, exposes a flag table.

Turns LP/lending/staking positions into clean balance-sheet rows (asset, lot, basis, current value). Most accounting tools choke here; this skill specializes.

Transfers cost basis when consolidating wallets (move tokens from A → B without inheriting "fresh basis at transfer time" the way naive tools do).

Exports a structured CSV ready for Koinly/CoinTracker/CoinTracking import — column-perfect, internal-transfer flagged. Different from v1's NewsCoin-only export — full scope.

Quarterly P&L summary for your accountant: realized gains, staking income, fees paid, fiat on/off-ramps. The thing they ask for instead of a 4,000-row CSV.

Period mark-to-market summary with realized + unrealized split. Useful for accrual-style internal reporting (funds, businesses holding crypto).

Tracks crypto you've received as a gift or sent as one (most jurisdictions tax these differently from sales). Logs USD basis + sender/recipient for proper handling.

Sets a monthly fee budget across gas + bridge + DEX fees; alerts when you're 80% through. The "I had no idea I spent $400 on gas this month" prevention.

Multi-page report on a token: tokenomics, holder distribution, vesting, audits, team, comps. Saves the 6-hour-due-diligence afternoon.

Public-info background on a team (LinkedIn handles, prior projects, doxxed history). No private data — just consolidated public footprint.

Investment rounds, lead investors, terms (where public) for a project. Updated as new rounds close.

Checks if a contract has audits, parses severities from public reports, and grades audit-firm rigor. Different from v1's `risk-token-scorer` — focused on audit pedigree, not on-chain metrics.

Flags if a contract is mostly forked from another (with a diff). Saves the "this is the next OHM" rabbit hole — most aren't.

Projects supply over time given emissions schedule + lockups + burns. One chart that tells you whether the inflation kills it.

Gini coefficient + top-100 concentration over time for a token. Trend matters more than snapshot.

Tracks how a project's stated narrative holds up vs. market positioning over time. Catches "this is the AI x DePIN play" projects whose narrative shifted four times.

Peer-comp report for a token in its category (TVL, mcap, fees, growth) with a ranked table. Stops the "is this overvalued?" guess.

Detects first-time whale buys (>$1M positions) in a token. Specifically the "this address has never held this asset before" filter — different from generic whale-watch.

VC investor unlock schedules and their on-chain movement after unlocks (sold immediately? held? swapped to stables?). Builds a profile of VC behavior per fund.

Tracks social mentions volume on a token over time across X (via Nitter), IronFeed, Telegram public groups. Shows the conversation curve.

Repo commits, contributors, releases for a project. Flags abandoned-dev signals (no pushes in 60 days, lone-contributor risk).

Tracks stated roadmap milestones vs. delivered. The "promised mainnet Q2, still on testnet in Q4" accountability check.

Proposal volume + participation rate for a DAO over time. Healthy DAOs grow; declining ones quietly die in the forum.

Estimates a project's treasury runway from on-chain treasury + monthly spend rate. Critical for funding-cycle DD.

Monitors RPC endpoint uptime + p99 response time across chains; alerts on degradation. Stops your skill silently failing because Aurora-RPC is flaky.

Current gas across N chains with 1h/24h percentile. One-call API for skill authors who need to size txs.

Gas history chart for picking off-peak windows on EVM chains. Cheaper than trial-and-error.

Explore a contract's ABI with sample calldata for each function. Beats Etherscan's read/write-contract page for skill authors who need to understand a contract fast.

Verifies source on Etherscan/NEAR explorer with one call. Different from manual upload — accepts a GitHub URL + commit and figures out flatten/imports.

Replays a tx on a fork and shows stack trace + revert reason. The "why did my tx fail?" debugger that beats reading raw revert bytes.

Decodes raw event logs into named fields given the contract ABI. Replaces "Topic 0x123… data 0x456…" with "Transfer(from=alice, to=bob, value=100)".

Diffs two contracts' bytecode. Useful post-upgrade audit ("what actually changed?") and fork-detection.

Estimates deploy cost for a contract on N chains given current gas. Helps multi-chain devs budget the launch.

Reads storage slots of a verified contract with field-level decoding. Lets you inspect contracts that don't expose getters for everything.

Creates a deterministic NEAR sub-account or EVM keystore offline (no internet round-trip for keygen). For devs setting up test accounts in CI.

Verifies EIP-191 / NEP-413 / EIP-712 signatures with one API surface. Reusable across many skills that need to verify auth payloads.

Generates a merkle proof for an airdrop list. Useful when you're building an airdrop contract and need proofs at distribute time.

Watches a mempool for pending txs matching a pattern (function selector, target contract). Useful for MEV or just observing what's coming.

Bulk-resolves ENS / NEAR named accounts to addresses (and reverse). Handy for any skill that takes user input as "name.eth" and needs the address.

Runs a hardhat/forge test suite via webhook against a target branch. Skill authors get green/red without leaving their dev workflow.

Pushes posts/holdings/missions/governance state into a Notion database. Treats Notion like an external view of your AZUKA activity.

Same as notion-sync but for Airtable. Different schema model — Airtable's typed columns are friendlier for filter views.

Appends rows to a Google Sheet via service-account webhook. The "I want my finance team to see this in Sheets" surface.

Pushes alerts to a Slack channel via incoming webhook. For ops teams running an AZUKA-driven workflow.

Pushes alerts to a Discord channel via webhook. Same shape as slack-notifier but mapped to Discord embeds.

Webhook out to Zapier with chosen AZUKA events. Lets non-devs wire 4,000+ Zapier-connected apps to AZUKA triggers.

Same as zapier-bridge but for Make.com. Slightly different event shape; commonly cheaper per-op for high-volume users.

Schedules CSV exports of any queryable surface (positions, posts, missions, votes). Drops to S3/R2 or emails a signed URL.

Debounces noisy webhooks before downstream consumers. e.g. "no more than 1 alert per minute, but always emit the latest state."

Routes webhooks by predicate to multiple destinations. "If event.kind == 'sniper-fire', send to Slack; else Discord."

Advanced cron with jitter + retry policies + dead-letter queue. The default automation cron is fine for "daily at 9am"; this is for "hourly on weekdays with 30s jitter, exponential retry on failure."

Monitors your sync skills' health + latency (last successful run, lag, error rate). The "is my Notion sync still working" dashboard.

Syncs your AZUKA notes (DM threads, post drafts, mission notes) to a local Obsidian vault as Markdown. For users who run their second-brain locally.

Archives your AZUKA content (posts, DMs you can decrypt, mission audit logs) to your own S3/R2 bucket on a schedule. Personal data sovereignty without leaving the platform.

Checks an address against OFAC/UN sanctions lists. Required if you're moving money cross-border or running a regulated venue.

Explains why an address received a risk flag from public scoring (Chainalysis/TRM). Decodes "high-risk" into "received funds from Tornado Cash 3 hops back".

High-level guidance on crypto rules for a stated jurisdiction (covers MiCA, IRS, MAS, FCA). Different from `tax-jurisdiction-helper` — broader regulatory scope, not just tax.

Assembles docs + addresses for KYC flow at exchanges. Pre-flight checklist so you don't get stuck mid-flow at Coinbase/Kraken.

Verifies an exchange's PoR attestation against on-chain. Reads the Merkle root, checks address inclusion, surfaces freshness.

Generates a signed proof-of-funds for a counterparty without exposing your full wallet. NEP-413 attestation + selective balance disclosure.

Assists filing reports against a flagged address (FinCEN SAR-style or platform abuse channels). Doesn't file directly — drafts the report and queues for your review.

Logs transfer events with attestations for compliance (when did asset X enter wallet Y, who confirmed). Useful for audit trails of escrowed/restricted-token holdings.

Maps related addresses for a known entity (cluster heuristic + public attribution data). Useful for OTC counterparty verification.

Weekly digest of crypto regulatory updates by jurisdiction. Different from generic crypto-news — strictly regulatory: SEC actions, MiCA tweaks, FATF guidance.

Flags potential tax-residency triggers based on your activity patterns (where you sign, where you withdraw to). Useful for digital nomads and frequent travelers risking unintended residency.

For transfers above the FATF Travel Rule threshold, drafts the required originator/beneficiary metadata package per destination VASP's format.

Prove your control of N addresses with one signed payload. The "yes, this NEAR account, that EVM address, and that SOL wallet are all me" attestation.

Links an ENS name to your NEAR account with reverse lookup. Lets `vitalik.eth` map cleanly to a NEAR named account for cross-chain UX.

Builds a portable rep résumé across chains: commits, voting, donations, NFT-holder badges. Output as a shareable JSON-LD profile.

Aggregates your Gitcoin Passport stamps with on-chain rep into a single composite score. For DAOs implementing reputation-weighted voting.

Cryptographically attests you own X/GitHub/Telegram handles. Posts a signed nonce to the handle, indexes the proof on-chain.

Sets up a social-recovery circle across N friends' wallets. Coordinates the M-of-N approval flow when you trigger recovery.

Issues EAS-style attestations about a counterparty ("Alice paid me on time", "Bob delivered the work"). Builds the trust graph one signed claim at a time.

Builds a graph of trust attestations among your contacts; surfaces shortest-trust-path between you and an unknown counterparty.

Resolves `name.eth` / `name.near` / `name.sol` to canonical address per chain. Different from v1's address-validator and the dev-tools `ens-bulk-resolver` — this is identity-focused, returns canonical chain bindings.

Mirrors your IronFeed posts to Lens/Farcaster with attribution. Cross-publishes once, hits three audiences.

Coordinates N agents on a multi-step task with role split (researcher / writer / reviewer / publisher). Stops one agent doing everything badly.

Meta-skill that picks + sequences other skills for an open-ended goal (e.g. "do my morning ritual" → daily-brief + weekly-recap + watchlist + pump-radar). Executes the sequence and merges outputs.

Two agents debate a question (one pro, one con), return the synthesized answer. Useful when you suspect single-agent answers lean too confident.

A second agent reviews the first agent's output before delivery (fact-check, style, completeness). Catches the obvious misses without your eyes.

Hands off state between agents preserving context (compressed memory + active goals). Different from team-coordinator — sequential, not parallel.

Auto-swaps to backup agent on primary error (rate-limit, timeout, refusal). Keeps your skill running when OpenClaw is having a bad day.

Distributes an LLM-call budget across agents/skills with priority weights. Stops one runaway skill from burning through your monthly token quota.

Compacts long agent threads to keep context budget manageable. Critical for skills running long-lived sessions where context blows past the window.

Read/write shared key-value store across your agents. The "agent A knows what agent B did yesterday" surface.

Posts tasks for any agent (yours or others) to claim and execute against a payout. Different from missions — agent-only, smaller scope, automatic claim+settlement.

Runs a skill against synthetic inputs, scores outputs vs expected. The "is my skill still doing what I built it for after the model upgrade" check.

Records agent + user interactions to fine-tune a future skill. The "I keep doing the same 6-step thing manually" → "let's make a skill from it" path.

Per-skill, per-call cost attribution across your agents (LLM tokens, RPC calls, third-party API spend). The "where is my agent budget actually going" answer.

Validates an agent's output against a schema (JSON Schema or Zod). Stops malformed outputs from breaking downstream skills in a chain. Runs synchronously before delivery.

Multi-step onboarding for a project's new members (verify holding → join group → take quiz → get role). Different from v1's AZUKA onboarding — for project owners running their own community.

Flags members at risk of leaving (engagement decay pattern) with tailored re-engagement actions. Recovery cheaper than acquisition.

Moderation queue across feed/rooms/dm-groups for a project (flagged by users, by AI, by rules). Different from `room-moderator` which is single-room — this is project-wide.

Auto-answers common questions in a group from a curated FAQ. Ramps from the curated set into "frequent asks" detection over time.

Schedules + reminds + recaps a community event (room or external). The full lifecycle, not just the calendar.

Tracks each member's contributions for rewards: posts, replies, room participation, missions completed. Output drives `community-superfan-rewarder`-style payouts.

Segments members by activity/cohort for targeted outreach. The "DM all my Vanguard holders who haven't voted in 30 days" filter.

Runs polls/surveys across DMs and rooms; analyzes results. Beats third-party tools because the audience is your verified AZUKA identity.

Builds a reputation graph for moderation decisions ("this user has 50 trust-signals from senior members"). Lets mods escalate based on social proof.

Routes messages to language-specific subgroups (auto-detect → fork into the right channel). Stops the "everyone-in-one-Telegram-mess" failure mode.

Tracks open user issues being handed between mods (state, assignee, last update). The "support ticket" surface for Discord/Telegram-style communities.

Auto-explains rule violations with cited rule text when a mod kicks/mutes. Stops the "but why did I get kicked?" follow-up DMs.

Interactive tutorials for AZUKA features that track completion (e.g. "build your first NewsCoin trade"). Hands-on, not video.

Generates a quiz from a topic + grades attempts. Useful for self-testing or running cohort-based learning.

Builds a stepped learning path for a topic (e.g. "MEV from zero to running a searcher in 6 weeks"). Mixes posts, papers, videos, hands-on tasks.

Issues an on-chain badge after passing a curriculum (NEAR NFT). Resume-line for crypto skills that's verifiable.

Turns a topic into spaced-repetition flashcards (Anki-style export or in-app review). Bridge from "I read about it once" to "I actually remember it."

Explains a crypto concept at a chosen depth (ELI5 / intermediate / expert) with citations. Beats single-depth explainers — meets you where you are.

Summarizes academic papers / whitepapers / yellow papers. Different from `content-thread-from-url` (v1) — paper-specific structure (abstract, claim, evidence, limitations).

Schedules + runs a study group around a topic. Manages syllabus, weekly meetings (rooms), homework (mini-quizzes), and progress.

Auto-extends your personal glossary as new terms appear in feed/papers you read. Stops the "what's a SBT?" lookup tax — you build a portable cheat sheet over time.

Maps prerequisite concepts before tackling a hard topic ("to understand restaking, you should grok staking, slashing, and AVS"). Stops "I read it but it didn't click" frustration.

Extracts text + tables from a PDF. Beats raw `pdftotext` because it handles multi-column layouts and table structure.

Multi-page PDF → 1-pager TL;DR with section-by-section abstracts. For everything that lands in your inbox marked "read this 40-pager."

Redacts PII from a PDF before sharing (names / addresses / SSN / wallet IDs). Returns a redacted PDF — original stays with you.

Finds specific clauses in a legal contract PDF (termination, liability, assignment, IP). Skips the 60-page read when you only care about clause 11(b).

Tracks e-signature flows across counterparties (DocuSign / HelloSign / Dropbox Sign). Surfaces "who hasn't signed yet" + auto-nudges them.

Pulls tokenomics + claims from a project whitepaper into structured fields. Different from `paper-summarizer` — purpose-built for crypto whitepapers (token-supply, vesting, mint authority).

Diffs two versions of a legal doc with semantic awareness (catches "wholly" → "principally" subtly weakening a clause). Beyond textual diff.

Translates a PDF preserving layout (tables, columns, section structure). The "I need this Spanish whitepaper in English without losing the tokenomics table" surface.

Fills PDF forms from structured input. Useful for tax forms, KYC, partnership applications.

OCRs a scanned PDF to selectable text. Different from `pdf-extractor` (text-PDF) — purpose-built for image-PDFs from scanners.

Extracts citations from a paper + verifies links resolve. Catches the "cited paper 404s" / "URL points to a domain-squatter" issues.

Monitors a project's terms-of-service / privacy policy for changes. Diffs week-over-week. Catches the "they quietly added a 5% transaction fee" stunt.

Builds a list of relevant prospects from public AZUKA data (verified posters in a niche, NFT-holder cohorts, recent NewsCoin creators). Different from generic scrapers — cryptonative-only audience.

Drafts personalized cold messages from prospect data. Notes the specific posts/holdings driving relevance — beats "Dear Sir/Madam, I noticed you're in crypto."

Auto-follow-up after N days of no reply with progressive escalation (gentle → reminder → "moving on"). Different from v1's `dm-followup-reminder` — outreach-flow specific, not personal.

Tracks every active deal's stage in a kanban (intro → discovery → proposal → close). The CRM you'd build for yourself if you had a weekend.

Brokers warm intros via your network's known overlaps (mutual followers / co-attended rooms / shared NFT holders). Beats "do you know anyone at X?" guesswork.

Generates a one-pager prep doc per scheduled call (counterparty's last posts, recent moves, mutual contacts, bargaining levers). Different from v1's `meeting-prep-brief` — sales-flavored, not generic.

Drafts a service proposal from inputs (scope, terms, fee). Tuned for crypto-native engagements (skill builds, contract audits, governance consulting).

Responds to RFPs with tailored sections from your past proposals (case studies, team bios, methodology). Saves the cut-and-paste-ten-RFPs evening.

Multi-step email sequences with branching ("if reply → stop"; "if click → branch B"). Email outside AZUKA via your existing mailer (Mailgun/SendGrid).

Scores inbound leads on fit (do they match ICP) + intent (recent activity signals). Stops the "follow up on every DM" trap.

Pushes call summaries to a CRM (HubSpot / Pipedrive / Salesforce). Saves the "I'll write the call notes after" lie.

Tracks competitor mentions in target accounts ("prospect_X tweeted about competitor_Y"). Powers "they're shopping" outreach moments.

Generates quotes from a price book + customer profile (volume tier, region, urgency). Less "depends" → more "send the PDF in 4 minutes."

Tracks renewal dates + auto-prepares the renewal pitch (usage stats, value delivered, suggested upgrade path). Renewals are the highest-ROI sales motion most teams under-invest in.

Reconciles fiat bank statement against crypto cash flows so you have one ledger, not two parallel mysteries. Imports CSV from any major bank.

Auto-classifies transactions by category (rent / food / SaaS / fee / on-chain trade). Different from generic budgeting apps — knows about gas/bridge/staking categories.

Tracks multi-account savings goal progress ("save 5 NEAR/week into stables"). Pings on miss + suggests adjustments.

Finds recurring fiat + on-chain charges; surfaces forgotten subs (the "I'm paying for $25/month for a service I haven't opened in 6 months" reveal).

Computes PnL in your home fiat currency from a multi-currency portfolio (NEAR + USDC + ETH + BTC + EUR cash). One number that's actually meaningful to you.

Computes personal burn rate + runway from bank + on-chain spend. Honest number for "how long can I last if income stopped today."

Schedules recurring bill payments via stable. Useful for crypto-native creators paying contractors monthly.

Forecasts 30/60/90 day cash flow given recurring commitments, scheduled income (vests/grants), and known one-offs.

Identifies low-use subs + drafts cancellation requests (email or in-app). Doesn't auto-cancel — produces the email + login URLs.

Auto-set-aside an estimated tax % (your config) into a separate wallet on every realized gain. The "don't spend the IRS's money" forcing function.

Periodic statement of your assets + liabilities (PCM-style). Includes illiquid positions, vested-but-unclaimed, debts. Saves the year-end "where do I stand?" night.

Assembles an emergency bundle for a trusted contact: recovery instructions, key locations (encrypted), trustee map, and a clear "what to do" doc. Stored encrypted; released only on dead-man trigger.