AgentCanary

Cross-asset market intelligence API for AI agents. 130+ endpoints across macro regime detection, risk scoring, trading signals (IGNITION/ACCUMULATION/DISTRIB...

Security Scan

The skill's declared purpose (market intelligence API) matches its instructions and it requests no elevated system access, but there are marketing/metadata inconsistencies and practical risks around payment and API key handling that a user should review before using.

Purpose & Capability

The SKILL.md describes an API-only market-intelligence service and the runtime instructions are limited to calling HTTPS endpoints under api.agentcanary.ai; there are no unrelated binaries, config paths, or environment variables requested. The resources the skill asks for (none) are proportionate to the described purpose.

Instruction Scope

Instructions stick to HTTP API calls and describe endpoints and polling patterns. Two practical concerns: (1) the service uses apiKey as a query parameter (apikey=YOUR_KEY) — the skill claims 'No secrets in prompt' but callers must ensure API keys are not injected into the LLM prompt or logs; how keys are supplied at runtime is not specified. (2) The onboarding requires making crypto deposits to a receiving address on external chains — depositing funds to an unfamiliar service carries financial and operational risk and is not something the agent can do on its own.

Install Mechanism

No install spec and no code files are present (instruction-only). Nothing will be written to disk by the skill itself; this is the lowest-risk install mechanism.

Credentials

The skill declares no required environment variables, no primary credential, and no config paths. That is consistent with an API-only skill that expects externally created API keys. There are no unexplained credential requests.

Persistence & Privilege

The skill is not set to always:true and uses normal autonomous invocation defaults. It does not request persistent system-level privileges or modify other skills' configs according to the provided metadata.

Assessment

This skill appears to be what it claims: an API-only market intelligence service. Before installing or using it, verify the operator and domain (visit agentcanary.ai from a trusted browser), and prefer depositing only a small test amount if you decide to fund an account. Confirm how your agent/runtime will store and send the apiKey so that the key is not exposed in LLM prompts, logs, or transcripts (use a secure secrets store or the platform's secure tool integration). The SKILL.md mentions a VirusTotal link and provides website links despite the registry metadata saying 'source: unknown' — consider asking the publisher for provenance (owner identity, company, or GitHub repo) and for API usage examples in your runtime environment before committing significant funds or automation.

Cross-asset market intelligence for AI agents. 130+ endpoints. Not raw data — intelligence.

Base URL: https://api.agentcanary.ai/api Auth: Wallet-based API keys. Create key → deposit USDC/USDT on any supported EVM chain → use key as query param. Briefs: 4× daily auto-generated intelligence — Radar (04:15), Signal (10:15), Pulse (16:15), Wrap (22:15 UTC). Telegram: @AgentCanary — live briefs. App: app.agentcanary.ai — dashboard, billing, API key management. Docs: api.agentcanary.ai/api/docs — interactive Swagger UI. Website: agentcanary.ai

---

Security

• API-only — HTTP GET/POST returning JSON. No local code, no binaries, no shell commands.
• No secrets in prompt — wallet-based auth. No API keys pass through the LLM context window.
• Read-only — fetches data. Cannot write, modify, or access your filesystem.
• No filesystem access — no file reads, no file writes, no directory listing.
• Security headers — Helmet.js (X-Powered-By removed, XSS protection, content-type sniffing prevention, strict transport security).
• Rate limiting — per-tier rate limits enforced server-side. Key creation: 5 attempts per 15 min per IP.
• Body size limit — 1MB max request body. Rejects oversized payloads.
• Error isolation — global uncaughtException/unhandledRejection handlers. Express error middleware. No stack traces in responses.
• Multi-chain deposits — USDC/USDT accepted on Base, Ethereum, Arbitrum, Optimism, Polygon. Same deposit address on all chains.
• VirusTotal verified — 0/62 detections.

---

Getting Started

1. POST /api/keys/create  { walletAddress: "0x..." }  → returns apiKey
2. Send USDC/USDT to the receiving address shown at agentcanary.ai (Base, Ethereum, Arbitrum, Optimism, Polygon)
3. POST /api/billing/check  { apiKey: "..." }  → auto-detects payment, credits account
4. Use endpoints:  GET /api/data/realtime-prices?apikey=YOUR_KEY

Minimum deposit: $5. Credits never expire. No subscriptions. No KYC.

---

Pricing

Tier	Cumulative Deposit	Per Call	Rate Limit	Access
Explorer	Free	$0.02	10/min, 50/day	Prices, news, whale alerts, Fear & Greed, regime
Builder	$50+	$0.02	60/min, 5K/day	+ macro, signals, calendar, volumes, newsletters
Signal	$150+	$0.015	120/min, 20K/day	All 130+ endpoints. AI reports. Orderbook. DeFi. Options.
Institutional	$500+	$0.01	300/min, unlimited	White-label. SLA. Custom integrations.

---

Default Agent Pattern

1. GET /api/macro/regime every 4–6 hours → classify risk environment
2. If Risk-Off → suppress trading, reduce exposure
3. If Risk-On → allow strategy execution, check signals
4. GET /api/data/whale-alerts every 15–30 min → event-driven interrupts
5. GET /api/signals/decision-engine before entries → multi-factor confirmation

AgentCanary is risk intelligence middleware. It tells your agent when conditions are favorable — your agent decides what to do.

---

Endpoint Categories

Full endpoint documentation with response examples: references/endpoints.md

Proprietary AC Endpoints (/api/...)

Category	Key Endpoints	Tier
Indicators (36)	/indicators, /indicators/summary, /indicators/:name, /indicators/:name/history — includes Bull Market Support Band, Pi Cycle, Wyckoff Structure, Stablecoin Composite, Composite Risk Score, and 31 more	Explorer–Signal
Scenarios	/scenarios/current, /scenarios/history, /scenarios/signals	Signal
Briefs	/briefs/latest, /briefs/feed, /briefs/archive, /briefs/:type	Explorer–Signal
Macro	/macro/regime, /macro/snapshot, /macro/signals, /macro/global-liquidity, /macro/us-m2, /macro/central-banks, /macro/supply-chain	Explorer–Builder
Regime	/regime, /regime/matrix, /regime/history	Signal
Signals	/signals/correlations, /signals/sector-rotation, /signals/btc-etf-flows, /signals/fear-greed, /signals/whale-alerts, /signals/geopolitical-risk, /signals/decision-engine + 26 more	Signal
Narratives	/narratives, /narratives/history, /narratives/:name	Signal
Expectations	/expectations, /expectations/rotation, /expectations/crowded, /expectations/early	Signal
DeFi	/defi/intelligence, /defi/pe-ratios, /defi/yields, /defi/perps, /defi/stablecoins, /defi/chains, /defi/unlocks, /defi/signals	Signal
BTC Options	/btc-options, /btc-options/maxpain, /btc-options/skew	Signal
Central Banks	/central-banks, /central-banks/balance-sheets, /central-banks/btc, /central-banks/stablecoins, /central-banks/gold, /central-banks/reserves, /central-banks/tic	Signal
Premiums	/premiums, /premiums/coinbase, /premiums/kimchi	Signal
Predictions	/predictions, /predictions/movers, /predictions/:slug	Signal
Sentiment	/sentiment/reddit	Signal
Mean Reversion	/mr/signals, /mr/trades, /mr/stats	Signal
Hindenburg	/hindenburg, /hindenburg/history	Signal
CAPE	/cape	Signal
Kill Conditions	/kill-conditions	Signal
Crypto Re-entry	/crypto-reentry, /crypto-reentry/history	Signal
Institutional	/institutional/13f	Signal
News	/news/trending, /news/stats, /news/market-analysis, /news/xtg-analysis	Signal

Data Endpoints (/api/data/...)

Cached datasets refreshed on schedule. 26 datasets covering prices, macro, crypto, news, institutional, calendar.

Dataset	Tier	Description
realtime-prices	Explorer	100+ crypto tokens, 24h change
yahoo-quotes	Builder	SPY, QQQ, VIX, TLT, DXY, Oil, 16 sector ETFs, stocks
whale-alerts	Explorer	Large crypto transactions
breaking-news	Explorer	Financial news with FinBERT sentiment
fear-greed	Explorer	Crypto Fear & Greed Index
macro-snapshot	Builder	30+ FRED series, regime, risk gauge, z-scores
funding-rates	Builder	Perpetual funding rates across exchanges
financial-calendar	Builder	High-impact economic events
newsletters	Builder	Curated newsletter intelligence
narrative-scores	Signal	21 narrative themes, crowding 1–5
btc-etf-flows	Signal	Bitcoin ETF daily flows
reddit-sentiment	Signal	14 subreddit sentiment analysis
decision-engine	Signal	Multi-factor crypto re-entry scoring
scenario-probs	Signal	6 macro scenario probabilities

Plus 12 more datasets. Full list in references/endpoints.md.

---

Signal Cadence Guide

Signal	Update Frequency	Recommended Cadence
Macro regime	Every 6h	Every 4–6 hours
Signal states (1d)	Daily close	Every 4–6 hours
Whale alerts	Real-time	Every 15–30 min
Funding rates	Every 8h	Every 4–8 hours
Breaking news	Real-time	Every 15–30 min
Briefs	4× daily	After each brief window
DeFi yields	Every 4h	Every 4–6 hours
BTC options	Daily	Every 4–6 hours

---

What It Does Not Do

• Does not predict prices — classifies regimes and states
• Does not place orders or replace execution logic
• Does not provide financial advice
• Does not guarantee returns

---

AgentCanary provides market data and intelligence for informational purposes only. Nothing constitutes financial advice.