What does this skill do, and when should you use it?
Whale Behavior Analyst is one skill inside AI-Market-Maker (AIMM), an open-source agentic hedge fund system, located at src/agents/operator/4.1_whale_behavior_analyst/. It relies on Nexus on-chain endpoints to monitor whale transactions and exchange flows, and answers direct questions like 'Are whales dumping BTC?' with a sell-pressure gauge, dump probability, and dry-powder alert. The skill is disabled by default in weights and is strictly opt-in. It ships as one of 14 skills bundled in the repository.
Exposes a Tier-0 contract analyze(ticker, market_data, nexus_context); supports direct queries such as 'Are whales dumping BTC?' or 'What's the exchange flow balance?'; serves /whale_behavior_analyst?ticker=BTC/USDT returning a sell pressure gauge, dump probability, and dry powder alert; sources data from Nexus on-chain endpoints (whale tx, exchange flow) and large transaction monitoring.
- A crypto trader who wants to verify whether whales are moving BTC to exchanges before opening a position
- A quant researcher integrating exchange net inflow/outflow as a signal factor into a strategy
- A risk team that wants a dump-probability alert when large on-chain transfers spike
- An AIMM operator who wants to opt into the whale-behavior analysis desk
How do you install this skill?
- Core function depends on paid Nexus on-chain endpoints (BNB Chain ecosystem); mainland-China reachability is unverified — confirm access before use.
- Dump-probability and related predictions carry no verification evidence; do not treat as sole basis for trading decisions — the agent is disabled by default with minimal weight.
- Static review only; no skill-specific tests found, failure modes and behavior when data is unavailable are unknown.
- Publisher is unverified by the FollowSkills registry; identity treated as unknown. AGPL-3.0 imposes copyleft obligations on integrators.
- Shell / CLI
- Network access
- Local filesystem
NexusDataClient on-chain endpointsPython 3.11+ (parent repo runtime)
The skill ships with the ai-market-maker repo: git clone https://github.com/olaxbt/ai-market-maker.git, then install dependencies per the README with uv sync --extra dev (TA-Lib must be installed separately). The skill file lives at src/agents/operator/4.1_whale_behavior_analyst/SKILL.md. OpenClaw users can run claw install https://github.com/olaxbt/ai-market-maker. Standalone install steps for this single skill are not documented.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/olaxbt/ai-market-maker.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/src/agents/operator/4.1_whale_behavior_analyst" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Are whales dumping BTC?
Enable the skill (note: disabled by default in weights, opt-in only — the exact enabling step is undocumented). Query /whale_behavior_analyst?ticker=BTC/USDT to get the sell pressure gauge, dump probability, and dry powder alert, or ask directly, e.g. 'Are whales dumping BTC?'. It depends on NexusDataClient on-chain endpoints, so Nexus data access must be available.
What are this skill's strengths and limitations?
- Direct natural-language query interface with concrete, readable outputs (gauge, probability, alert)
- Grounded in objective on-chain data — large transactions and exchange flows — rather than sentiment
- Follows the unified Tier-0 Input→Output contract, easy to plug into the AIMM workflow
- Disabled by default, so it does not affect users who don't opt in
- Hard dependency on Nexus on-chain endpoints; unusable without Nexus access
- Opt-in by default and the enabling procedure is not documented in the source
- No skill-specific tests, accuracy metrics, or backtest evidence
- Covers only the whale-behavior dimension and cannot be a standalone trading decision
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Whale Behavior Analyst (4.1) this page | 44 · Not recommended | ★ 2.1k | 2mo ago | AGPL-3.0 |
| News Narrative Miner | 46 · Use with care | ★ 2.1k | 2mo ago | AGPL-3.0 |
| BTC Bottom Model (Bitcoin Market Heat Scoring) | 49 · Use with care | ★ 1.1k | 6mo ago | MIT |
| Pro Bias Analyst | 44 · Not recommended | ★ 2.1k | 2mo ago | AGPL-3.0 |
| Liquidity Order Flow Analyst (4.2) | 47 · Use with care | ★ 2.1k | 2mo ago | AGPL-3.0 |
How did FollowSkills review this skill?
Read-only on-chain whale analysis, disabled by default (AGENT_WEIGHTS_DEFAULT=0.05, opt-in), no destructive side effects; deductions: depends on external paid Nexus endpoints, no data provenance attribution, no fallback/rollback path, no permission scoping or confirmation mechanism at the skill level despite repo-level SECURITY.md.
SKILL.md and persona.md are internally consistent (tier0/v1 JSON contract, threshold rule Dump≥0.65→bear vote); deductions: no skill-specific implementation or tests visible (whale key path absent from shown suites), NexusDataClient dependency unauditable, failure feedback on abnormal input undefined.
Clear scenario (on-chain trader querying whale sell pressure/exchange flow), explicit trigger queries, non-default scope disclosed; deductions: no Chinese support, core function depends on Nexus on-chain APIs (BNB Chain ecosystem, mainland-China reachability unverified), limited boundary/non-fit documentation.
Well-layered docs (SKILL.md + persona.md + few-shot + output schema), clear AGPL-3.0 license; deductions: no per-skill versioning/changelog, maintenance ownership and update path not stated at skill level, publisher unverified.
Output format (gauge/probability/alert) directly usable, few-shot shows expected behavior; deductions: correctness and marginal value of dump-probability estimates entirely unverified, skill is disabled by default indicating authors excluded it from the default research combo.
Only few-shot examples and a stray dump_probability field in repo fixtures; deductions: no third-party execution evidence, no reproducible representative outputs, and the README's backtest return table has no direct linkage to this skill.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →