agentleFS
dasein108dasein108/ai-crypto-scoring/.claude/skills/token-research/SKILL.md

token-research

Skill0 starsChanged 5 months agoMITView on GitHub

On-chain analysis of an EVM crypto token — risk, structural quality (6-pillar score), liquidity, holders, locks/unlocks, and full dossiers — by driving the token-research CLI. Use when the user wants to research, score, or risk-assess a specific EVM token (Ethereum/Base/Arbitrum/Optimism) by ticker, project name, or contract address, or compare several EVM tokens. For non-EVM L1s (HBAR, SOL, ATOM), pre-TGE programs, or purely qualitative trust/DD, defer to the cpd-crypto-analysis skill instead.

  • Installs packages

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. DefiLlama/defillama-skills★ 32same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repositoryOne other file it gives its agents.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 0 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run with the console script token-research <command> …, or without installing via PYTHONPATH=src python3 -m token_research <command> ….

    line 34
  2. Very specific:

    Always pass --address and --chain when known — it pins the exact contract and skips resolver guesswork (squatter-token defense).

    line 36
  3. Very specific:

    Read the output structure: scores (pillars + composite + coverage_ratio), narrative, evidence_summary, per-module metrics, warnings, …

    line 59
  4. Very specific:

    When a pillar surprises you, drill in with the targeted command: compare (effective circulation / protocol-owned vs EOA), holders …

    line 61
  5. Very specific:

    Every command returns a CommandResult: metrics, sources, warnings, coverage_gaps, resolved_identity, generated_at.

    line 75

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

token-research (CLI-driven EVM token analysis)

Drive the token-research CLI to produce on-chain-grounded, provenance-tracked token analysis. The CLI is the engine; this skill routes the request to the right command(s) and reads back the structured output.

Step 0: Route first (data regime)

Apply this BEFORE running anything:

EVM token with a known or derivable address (Ethereum / Base / Arbitrum / Optimism)?
 ├─ yes → use this skill (below)
 └─ no  → STOP. Hand off to the `cpd-crypto-analysis` skill:
            • non-EVM L1 (HBAR, SOL, ATOM, TIA, TON, native BTC…)
            • pre-TGE / no token yet (e.g. a points program)
            • a purely qualitative trust / team / governance question

The CLI is EVM-first and token-centric. Its refusals are the hand-off signal, not a failure. See docs/skills-architecture.md.

Setup (once)

pip install -e .[dev]            # installs the `token-research` console script
# optional, for momentum/portfolio price features:
pip install -e ".[ccxt]"

Run with the console script token-research <command> …, or without installing via PYTHONPATH=src python3 -m token_research <command> …. Add TOKEN_RESEARCH_OFFLINE=1 to skip all network calls (deterministic, for smoke checks).

Always pass --address and --chain when known — it pins the exact contract and skips resolver guesswork (squatter-token defense). Default output is JSON; deep-report also supports --format markdown.

Modes (pick by purpose × depth)

Mode: quick — triage in seconds

For "is this worth a closer look?" Run the two scoring systems and read the headlines:

token-research score <TOKEN> --chain <CHAIN> --address <ADDR>   # 6-pillar structural quality 0–100
token-research risk  <TOKEN> --chain <CHAIN> --address <ADDR>   # CPD 2.0 risk %/tier

Report: composite score + the weakest pillar, risk tier, and any warnings / coverage_gaps. A high score with many coverage gaps is a weak read — say so.

Mode: deep — full dossier

For a real write-up. One command runs every collector + the 6-pillar score + narrative:

token-research deep-report <TOKEN> --chain <CHAIN> --address <ADDR> --format markdown

Read the output structure: scores (pillars + composite + coverage_ratio), narrative, evidence_summary, per-module metrics, warnings, coverage_gaps. Persisted to .token-research/reports/ unless --no-persist.

When a pillar surprises you, drill in with the targeted command: compare (effective circulation / protocol-owned vs EOA), holders (concentration, gini, nakamoto), liquidity (depth, LP locks, slippage), unlocks (cliff overhang; add --with-premium for curated Tokenomist data), locks / staking (custody fingerprinting), yields (TVL, fees, revenue model).

Mode: compare — several tokens side by side

The compare command is a single-token effective-circulation view, so for head-to-head run the same command per token and diff:

for each token: token-research score <TOKEN> --chain <CHAIN> --address <ADDR>

Present a table: composite, the driving pillars, and risk tier per token. Reconcile differences — two tokens with the same composite can have very different pillar shapes (see docs/walkthrough.md).

Reading the output (the contract)

Every command returns a CommandResult: metrics, sources, warnings, coverage_gaps, resolved_identity, generated_at. Three rules:

  1. Trace, don't trust — every metric carries its sources (on-chain RPC > explorer > vendor).
  2. Coverage gaps are signal — they list what couldn't be answered and how to fill it; the composite is penalized for them. Surface them, never hide them.
  3. Pillars over composite — report which pillar is weak, not just the number. Pillar definitions, thresholds, and market-dependency: docs/scoring-overview.md and articles/01-estimation-methodology.md.

Compose with CPD for a complete report

For a full "is this a good, trustworthy token?" answer, run this skill for the on-chain numbers, then the cpd-crypto-analysis skill for the sourced trust layer, and merge. The CLI risk and the CPD score share one source of truth (docs/risk_revenue_estimation.md); where they diverge, that divergence is itself a finding worth reporting.

Reference

  • Command surface + flags + output schema: docs/cli-reference.md
  • Scoring systems (score / risk / signal / fairlaunch): docs/scoring-overview.md
  • End-to-end walkthrough: docs/walkthrough.md
  • Skill routing rationale: docs/skills-architecture.md

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.