Token Analysis Skill

Founder-first research framework for early-stage crypto tokens. Evaluate tokens systematically, manage a watchlist, and monitor positions.

For deep analysis, read references/playbook.md in this skill's directory — it contains the investment philosophy (metagame lifecycle, attention theory, probabilistic thinking) that informs every evaluation.

For a completed example showing expected depth and tone, see references/example-analysis.md.

Analysis Framework

Always verify token identity by contract address + chain before analysis. Ticker symbols are not unique — multiple tokens can share the same symbol. Use the contract address from DexScreener as the canonical identifier.

When evaluating any token, assess these in order:

1. Founder / Dev Deep Dive (MOST IMPORTANT)

• Find the dev/founder's X/Twitter account. Search for their recent tweets, bio, follower count.
• Background: prior projects, exits, credentials.
• Social graph: who follows them? Notable backers/mutuals? Quality of engaged accounts (builders vs bots).
• Activity: posting frequency, shipping updates, engagement levels.
• Red flags: anonymous with no history, inactive, only posting price action.
• If no founder/dev can be identified, flag this as a significant risk factor. Anonymous with no history is a red flag — not a dealbreaker if product traction is strong, but requires higher conviction on other criteria.
• Always hyperlink: [@dev](https://x.com/dev)

2. Product Reality

• Does a working product exist or is it vaporware?
• Onchain activity beyond token trading?
• Users, integrations, traction metrics?

3. Team Signal

• Other team members, advisors, backers.
• Background (pseudonymous is fine if building; silent is not).
• Hiring? (signal of ambition and runway)

4. Market Structure

• FDV vs comparable projects (find the comp, calculate the gap).
• Buy/sell ratio and trend.
• Liquidity depth relative to FDV.
• Holder distribution (whales, concentration).
• Token age and price history.

5. Narrative Fit

• Does it ride a live narrative? (AI agents, DeFi, L2 ecosystem, etc.)
• Who's talking about it? (quality > quantity)
• Is discourse growing or fading?

6. Decision Output

Always end with:

• Verdict: watch / small entry / conviction entry / pass
• Entry target: specific FDV or price
• Kill conditions: what makes this trade dead
• Catalyst: what would make you scale in

Output Template

Use this format for every analysis:

## [SYMBOL] Analysis — [DATE]

**Token:** [name] ([symbol])
**Address:** [contract address](https://dexscreener.com/[chain]/[contract address])
**Chain:** [chain]
**FDV:** $[X] | **Liq:** $[X] | **Age:** [X days]

### Founder / Dev
[findings — link all X accounts as [@handle](https://x.com/handle)]

### Product
[findings]

### Team
[findings — link all X accounts as [@handle](https://x.com/handle)]

### Market Structure
[findings — include FDV vs comps, buy/sell ratio, liq/FDV ratio, holder concentration]

### Narrative
[findings — attention state, discourse trend, who's talking]

### Verdict
- **Decision:** watch / small entry / conviction entry / pass
- **Entry target:** [specific FDV or price]
- **Kill conditions:** [what makes this dead]
- **Catalyst:** [what would make you scale in]
- **Confidence:** [low / medium / high]

### Sources
- [list URLs used: DexScreener, X posts, etc.]

Formatting rules:

• Contract addresses must hyperlink to DexScreener: [0x...](https://dexscreener.com/base/0x...)
• X/Twitter accounts must hyperlink: [@handle](https://x.com/handle)
• Never use bare addresses or @handles without links

Data Sources

DexScreener (primary — free, no API key)

Search by token:

https://api.dexscreener.com/latest/dex/search?q=SYMBOL

Search by address:

https://api.dexscreener.com/latest/dex/tokens/ADDRESS

Web UI: https://dexscreener.com/search?q=SYMBOL_OR_ADDRESS

X/Twitter Research (X API v2)

Social discourse is critical for early-stage tokens. Use the X API v2 to search for relevant posts. Requires an X API bearer token ($X_BEARER_TOKEN).

Search for the token:

curl -s "https://api.x.com/2/tweets/search/recent?query=SYMBOL%20-is:retweet&max_results=20&tweet.fields=created_at,public_metrics,author_id&expansions=author_id&user.fields=username,public_metrics" \
  -H "Authorization: Bearer $X_BEARER_TOKEN"

Search for dev activity:

curl -s "https://api.x.com/2/tweets/search/recent?query=from:devhandle%20-is:retweet&max_results=10&tweet.fields=created_at,public_metrics" \
  -H "Authorization: Bearer $X_BEARER_TOKEN"

Look up a user profile:

curl -s "https://api.x.com/2/users/by/username/handle?user.fields=description,public_metrics,created_at" \
  -H "Authorization: Bearer $X_BEARER_TOKEN"

What to search for:

• The token name/symbol — gauge discourse and sentiment
• The dev/founder handle (from:devhandle) — check activity and shipping
• Notable accounts mentioning the token — quality of attention

If X API is unavailable: Fall back to web search for recent X posts, or check the dev's public X profile directly at https://x.com/handle. The analysis framework works with any source of social data — the API just makes it systematic.

Chain-Specific Gotchas

The framework is chain-agnostic, but watch for these:

• Base / L2s: Bridge liquidity can be thin — check if liquidity is native or bridged. Low gas means more bot activity; buy/sell ratios can be misleading.
• Solana: Token extensions can hide transfer fees or freeze authority. Always check mint authority and whether it's revoked. pump.fun tokens have standardized mechanics but graduation to Raydium changes the liquidity profile.
• Ethereum L1: Gas costs mean smaller traders are priced out — holder base skews whale-heavy. Check if the token is also deployed on L2s.
• General: Always verify the contract address against DexScreener. Check for honeypot indicators: can you actually sell? Is there a max transaction limit? Is the deployer wallet still holding a large share?

Watchlist Management (Optional)

Optional persistent tracking. Skip this section if you just need one-off analysis.

Watchlist file: watchlist.json in the skill directory.

Schema

{
  "tokens": [
    {
      "symbol": "EXAMPLE",
      "address": "0x...",
      "chain": "base",
      "addedAt": "2026-01-15",
      "status": "watching",
      "thesis": "One-line investment thesis",
      "fdvAtAdd": 1000000,
      "entryTargets": [
        { "fdv": 500000, "note": "dip buy if thesis intact" }
      ],
      "scaleTargets": [
        { "fdv": 2000000, "condition": "user growth >1000" }
      ],
      "killConditions": [
        "dev goes quiet >3 days",
        "liquidity drops below $200K"
      ],
      "exitTargets": [
        { "fdv": 10000000, "action": "take 50% off" }
      ],
      "notes": [],
      "lastChecked": null
    }
  ]
}

How to Use

These are example phrases you say to your agent — not literal CLI commands.

Add token — say something like "watch [token]" or "add [token] to watchlist":

1. Fetch DexScreener data (price, FDV, liquidity, age)
2. Search X/Twitter for the token and dev
3. Build thesis, entry targets, kill conditions
4. Add to watchlist.json

Check watchlist — say "check watchlist" or "watchlist status":

1. For each token with status "watching" or "entered":
   • Fetch current data from DexScreener
   • Compare to entry/exit/kill targets
   • Flag anything that hit targets
2. Output concise status table

Analyze token — say "analyze [token]" or paste a DexScreener link:

1. Fetch DexScreener data
2. Search X/Twitter for token and dev
3. Run the 6-step analysis framework above
4. Output decision with sizing recommendation

Remove token — say "remove [token] from watchlist":

1. Set status to "killed" or "exited" with reason
2. Keep the entry for track record (don't delete)

Token Monitoring (Optional)

This section describes the monitoring pattern. Implementation depends on your agent platform (OpenClaw cron, scheduled tasks, etc.).

Hourly monitoring pattern

For each monitored token, the monitoring job should:

1. Fetch DexScreener data (price, FDV, liquidity, volume, buy/sell counts)
2. Search X/Twitter for dev activity and new mentions
3. Compare to previous check — flag significant moves
4. Alert on: price ±15% in 1h, dev posts major update, kill condition triggered, entry target hit, liquidity change ±20%

Monitor log format

Keep a log file per token at monitors/SYMBOL.md:

# SYMBOL Monitor Log
## Token Info
- Address: 0x...
- Chain: base
- Dev: @handle
- Started: YYYY-MM-DD

## Updates
### YYYY-MM-DD HH:MM UTC
- Price: $X | FDV: $X | Liq: $X
- 1h: X% | 24h: X%
- Volume 1h: $X | Buys: X | Sells: X
- Dev activity: [summary or "none"]

Principles

• Express conviction through sizing, not certainty.
• A "pass" is a valid and good outcome — there are infinite opportunities.
• Always show your work: scenarios, probabilities, expected value.
• If using the watchlist, update watchlist.json after every analysis.

Disclaimer

This skill is experimental. It provides a research framework — not financial advice.

• Outputs are indicative and have not been backtested against historical performance.
• No analysis produced by this skill should be treated as a recommendation to buy or sell any asset.
• Early-stage tokens are highly volatile and carry significant risk of total loss.
• Always do your own research. An AI agent following a framework is not a substitute for human judgment.
• The authors accept no liability for any losses incurred from acting on outputs of this skill.

---

_{Built by Fair}