solana-token-research
Research a Solana token or daily hot-token set using market structure, liquidity and AMM reserves, directional wallet flows, source-of-funds tracing, holder normalization, MEV/bot filtering, KOL evide
SKILL.mdskills/solana-token-research/SKILL.md
GitHubSolana Token Research
Produce an evidence-backed snapshot before giving a narrative or trade view.
Start
- Resolve the exact mint address and reject symbol-only identification.
- Record the snapshot timestamp, timezone, data source, pair address, quote asset and pool venue.
- Use GMGN market/token/holder/portfolio/track capabilities when available. Supplement gaps with Solana RPC, DEX pool data and current public sources.
- Read references/evidence-framework.md before attributing wallets, market makers, insiders or KOLs.
- Read references/report-schema.md when producing a ranking, spreadsheet, monitor or publishable report.
Research workflow
1. Market structure
- Collect price, market cap, FDV, liquidity, holders and volume for 5m, 15m, 1h, 6h and 24h windows when available.
- Compare liquidity growth with price and volume growth. Flag price expansion without matching liquidity.
- Split the move into accumulation, markup, acceleration, distribution and stabilization phases when the chart supports them.
- Identify support, reclaim, confirmation, prior-high and invalidation levels.
2. Pool and execution evidence
- Inspect the dominant pool, its token/SOL reserves, liquidity concentration and reserve changes across the event.
- Estimate the quote-side capital required for the observed price displacement using the AMM curve. State assumptions.
- Distinguish gross turnover from directional net flow.
- Flag shallow liquidity, concentrated LP ownership, removable liquidity and abnormal slippage.
3. Wallet flows
- Rank net buyers and net sellers over the event window; include trade count, amount, average entry, remaining balance and realized behavior.
- Trace the immediate source of funds. Classify it as CEX withdrawal, bridge inflow, internal token rotation, wallet transfer or unknown.
- Score wallet quality from repeated history, not one well-timed trade.
- Separate directional holders from sandwich bots, bundlers, arbitrageurs and same-block round trips.
- Treat market-maker attribution as a hypothesis unless repeated two-sided quoting, inventory behavior and counterparty structure support it.
4. Holder structure
- Normalize DEX vaults, pool authorities, CEX custody, bridges, burn addresses and program accounts before calculating private concentration.
- Report adjusted private Top 10, related-wallet clusters, creator/dev holdings, profitable supply, trapped supply and holder-duration distribution.
- Compare price direction with holder-count direction. Explain whether the pattern suggests accumulation, dispersion or noise.
5. Narrative and social evidence
- Search current X/Twitter and official project channels when the move may be news-driven.
- Map a KOL only with public identity evidence or a reliable labeled-wallet source.
- Explain why the account matters: audience, prior calls, project role, wallet performance or information access.
- Separate confirmed announcements, credible rumors and engagement-only chatter.
6. Decision model
Score 0–100 using:
- Market and liquidity quality: 20
- Directional capital quality: 25
- Flow persistence: 15
- Adjusted holder structure: 15
- Narrative and social confirmation: 10
- Execution safety: 15
Cap the score at 45 for critical unresolved risks such as unverifiable mint identity, removable liquidity, extreme private concentration, active creator dumping or predominantly bot-generated volume.
Do not output a binary buy call alone. Provide:
- base case, bull case and bear case
- entry confirmation conditions
- invalidation conditions
- monitoring metrics and alert thresholds
- evidence that would change the conclusion
Output rules
- Lead with the conclusion and confidence level.
- Label facts as observed, labeled by a provider, or inferred.
- Include full mint and relevant wallet addresses in a copyable appendix.
- Include timestamps and methodology caveats.
- Never expose API keys, cookies, private keys or unredacted personal data.
- Do not execute trades unless the user separately and explicitly authorizes financial execution.
- State that the research is not investment advice.