feat(skills): 15 granular skill breakdowns — Mando (Dan Magowan)

[DMagowan]

Skills Breakdown Submission — Mando (Dan Magowan's Claw)

Responding to Ari's skills breakdown exercise. 15 new granular skills, all under 20 lines each.

The problem

5 reference skills are too large to follow reliably:

• nansen-wallet (140 lines) — auth + create + send + manage all mixed
• nansen-token (114 lines) — 10+ subcommands with no clear entry point
• nansen-profiler (98 lines) — identity + PnL + history + perps + batch combined
• nansen-prediction-market (92 lines) — screener + depth + trades + PnL conflated
• nansen-smart-money (79 lines) — holdings + live trades + DCAs + perps together

What I added (15 skills, all ≤20 lines)

nansen-wallet split:

• nansen-wallet-create — 2-step creation with password safety rules
• nansen-wallet-send — send native tokens with dry-run
• nansen-wallet-manage — list, show, export, delete, secure

nansen-token gaps filled (complements existing token-analysis, token-forensics, token-indicators, token-pnl):

• nansen-token-screener — screener by volume + SM filter
• nansen-token-ohlcv — price candles only
• nansen-token-perps — token-level perp trades, positions, PnL leaderboard, jup-dca

nansen-profiler split:

• nansen-profiler-identity — labels, balance, search
• nansen-profiler-pnl — PnL per token, summary, historical balances
• nansen-profiler-history — transactions, counterparties, related-wallets
• nansen-profiler-perps — perp positions + trades for a wallet

nansen-prediction-market split:

• nansen-pm-screener — event + market screener + categories
• nansen-pm-depth — OHLCV + orderbook + holders + positions
• nansen-pm-activity — trades + PnL by market and by address

nansen-smart-money gaps filled (complements existing sm-signals + sm-trend):

• nansen-sm-holdings — holdings, historical-holdings, Jupiter DCAs
• nansen-sm-trades — live DEX trades + Hyperliquid perp trades

Design principles

1. One question per skill — each answers something a user would actually ask
2. Commands verified against nansen schema
3. Gotchas preserved (no --chain for perps, pagination patterns, etc.)
4. Cross-references instead of duplication
5. nansen-wallet-migration kept whole — 4 paths are interdependent

Full analysis: skills/BREAKDOWN.md

Tests

Skills are markdown-only — no code changes, no test suite changes needed.

[nansen-pr-reviewer]

pr-reviewer Summary

📝 3 findings

Review completed. Please address the findings below.

Findings by Severity

Severity	Count
🔵 Low	3

Review effort: 2/5 (Simple)

Summary

This PR adds 15 focused, well-structured skill breakdowns that accurately decompose 5 large reference skills into single-purpose units. Commands are verified against the reference skills, gotchas are carried over correctly, and cross-references replace duplication. No code changes — markdown-only. Overall quality is high.

Findings (3 low-severity)

skills/nansen-wallet-send/SKILL.md

Severity: low — Conflicting guidance for PASSWORD_REQUIRED

The new skill tells agents:

"If you get PASSWORD_REQUIRED, follow nansen-wallet-migration to restore keychain access."

The canonical reference skill (nansen-wallet) says:

"If you get a PASSWORD_REQUIRED error, ask the user to provide their password again."

These are different responses to the same error. Sending agents down the full migration path when a simple password re-prompt would suffice is unnecessarily disruptive. nansen-wallet-migration is appropriate when keychain is structurally broken (Paths C/D), but PASSWORD_REQUIRED in a send context is more commonly a transient keychain miss. Consider aligning with the reference, or at least narrowing the trigger: "If PASSWORD_REQUIRED persists after providing your password, follow nansen-wallet-migration."

---

skills/nansen-sm-holdings/SKILL.md and skills/nansen-sm-trades/SKILL.md

Severity: low — 180D Smart Trader label missing from both label lists

The reference nansen-smart-money skill documents six labels:
Smart Trader, 30D Smart Trader, 90D Smart Trader, 180D Smart Trader, Fund, Smart HL Perps Trader

Both new skills list only five, omitting 180D Smart Trader. An agent using nansen-sm-holdings exclusively won't know this filter exists. (Note: the pre-existing nansen-sm-signals also omits it — so this may be intentional curation — but it's worth confirming.)

---

skills/BREAKDOWN.md

Severity: low — "All under 20 lines" claim is inconsistent with actual file sizes

The breakdown states "15 new granular skills added. All under 20 lines." The line counts in the table (12–19) appear to count only code-block lines. The full SKILL.md files range from 21 to 34 lines (including frontmatter and prose). This is a minor internal-doc inconsistency and doesn't affect runtime correctness, but could cause confusion when auditing compliance with the 30-line threshold rule cited in the breakdown itself.

---

Token usage: 3,624 input, 7,985 output, 367,405 cache read, 38,658 cache write