Install
$ agentstack add skill-tradeparadex-paradex-skills-vault-intelligence ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
Paradex Vault Intelligence
Turns raw vault data from Paradex MCP into investment-grade vault analytics. Helps users discover, compare, and monitor vaults for passive income strategies.
Available MCP Tools (data sources)
| Tool | What it gives you | Key params | |---|---|---| | paradex_vaults | Vault config, owner, status, kind | vaultaddress, jmespathfilter | | paradex_vault_summary | Performance metrics (ROI, PnL, drawdown, volume, TVL) | vaultaddress, jmespathfilter | | paradex_vault_positions | Current open positions | vaultaddress | | paradex_vault_balance | Available/locked/total balance | vaultaddress | | paradex_vault_account_summary | Account health, margin, exposure | vaultaddress | | paradex_vault_transfers | Deposit/withdrawal history | vaultaddress |
Capabilities
1. Vault Discovery & Screening
Use paradex_vault_summary with JMESPath to screen the full vault universe:
By performance:
# Top 5 by total ROI
"sort_by([*], &to_number(total_roi))[-5:]"
# Profitable in last 7 days
"[?to_number(roi_7d) > `0`]"
# Best 30-day performers
"sort_by([*], &to_number(roi_30d))[-10:]"
By risk:
# Lowest max drawdown (safer vaults)
"sort_by([*], &to_number(max_drawdown))[:5]"
# Low recent drawdown + positive returns
"[?to_number(max_drawdown_30d) `0`]"
By size/activity:
# Largest by TVL
"reverse(sort_by([*], &to_number(tvl)))"
# Most active by 24h volume
"reverse(sort_by([*], &to_number(volume_24h)))"
# Most depositors (social proof) — apply to paradex_vaults only, NOT paradex_vault_summary
# paradex_vault_summary does not include num_depositors; use paradex_vaults for this filter
"reverse(sort_by([*], &to_number(num_depositors)))"
Note: num_depositors is available on paradex_vaults, not paradex_vault_summary. For depositor-count screening, call paradex_vaults and join on vault_address to attach performance metrics from paradex_vault_summary.
2. Vault Deep Dive
For a specific vault, gather comprehensive data:
paradex_vaults— get config: owner, kind, status, creation detailsparadex_vault_summary— performance: ROI (24h/7d/30d/total), PnL, drawdowns, volume, TVL, token priceparadex_vault_positions— current holdings: which markets, sizes, unrealized PnLparadex_vault_account_summary— account health: margin usage, leverage, exposureparadex_vault_balance— cash position: available vs. lockedparadex_vault_transfers— fund flows: deposit/withdrawal patterns
3. Vault Comparison
When comparing 2+ vaults, build a comparison matrix:
| Metric | Vault A | Vault B | Better | |---|---|---|---| | Total ROI | X% | Y% | ✓ higher | | 30d ROI | X% | Y% | ✓ higher | | Max Drawdown | X% | Y% | ✓ lower | | Sharpe-like ratio | ROI/DD | ROI/DD | ✓ higher | | TVL | $X | $Y | context-dependent | | # Depositors | N | M | ✓ higher (social proof) | | 24h Volume | $X | $Y | ✓ higher (activity) | | Position count | N | M | context-dependent |
Risk-adjusted ranking: Compute a simple Sharpe-like ratio: total_roi / max_drawdown (higher = better risk-adjusted returns). For any vault where the ratio is unreliable — this includes vaults with zero or near-zero drawdown, fewer than 30 days of data, and small-TVL vaults where ROI may be inflated by small base effects — use total_roi alone but always output the exact phrase: "insufficient drawdown history — ratio unreliable". Use this exact language in both comparison tables (as a footnote †) and in any narrative text. Do not substitute synonyms ("limited history", "small TVL", etc.).
4. Vault Risk Assessment
For each vault, assess and report:
- Drawdown risk: maxdrawdown vs. maxdrawdown_30d — is drawdown getting worse?
- Concentration risk: from positions — how many markets, what % of exposure is in largest position?
- Leverage risk: from account_summary — current leverage vs. available margin
- Liquidity risk: from TVL + transfers — is TVL stable, growing, or declining?
- Operator activity: from volume — is the vault actively trading or dormant?
Risk score (1-5):
- 1 (Low): Low drawdown, diversified positions, moderate leverage, stable/growing TVL
- 3 (Medium): Some drawdown history, concentrated in 2-3 markets, moderate leverage
- 5 (High): Large drawdowns, single-market concentration, high leverage, declining TVL
5. Vault Monitoring & Alerts
When asked to monitor a vault, check for:
- ROI dropping below a threshold
- Drawdown exceeding user-defined limit
- TVL declining (depositor exodus)
- Position concentration changing significantly
- New large positions opened (strategy shift)
6. Vault Recommendation Engine
When a user asks "which vault should I deposit in?", gather their preferences:
Risk tolerance:
- Conservative: prioritize low drawdown, stable returns, high TVL, many depositors
- Moderate: balance ROI and drawdown, accept some concentration
- Aggressive: prioritize highest ROI, accept higher drawdown and concentration
Time horizon:
- Short-term: weight roi24h and roi7d more heavily
- Medium-term: weight roi30d and lastmonth_return
- Long-term: weight totalroi and maxdrawdown
Then screen, score, and present top 3-5 vaults with reasoning.
Output Format
Vault Screening Results
Screening format = list only. Do NOT add per-vault narrative sections or deep-dives. One bullet block per vault, then done. Reserve deep-dives for explicit "analyze vault X" requests.
## Paradex Vault Screening — [criteria]
Found N vaults matching criteria. Top picks:
| # | Vault | Total ROI | 7d ROI | Max DD | Sharpe-like | TVL | Risk |
|---|---|---|---|---|---|---|---|
| 1 | [Name/Address] | X% | X% | X% | X.Xx | $X | X/5 |
| 2 | [Name/Address] | X% | X% | N/A† | N/A† | $X | Unrated† |
| 3 | [Name/Address] | X% | X% | X% | X.Xx | $X | X/5 |
† Insufficient drawdown history (<30 days data) — Sharpe-like ratio unreliable; total ROI shown only.
*Past performance does not guarantee future results. Verify figures on the Paradex vault page before depositing. Vault operator strategies are not fully transparent — positions show current holdings but do not reveal the full strategy or risk model.*
Vault Deep Dive
## Vault Analysis — [address]
### Performance
[ROI table across timeframes]
### Current Positions
[Position breakdown with market, size, unrealized PnL]
### Account Health
[Margin usage, leverage, available capacity]
### Risk Assessment
[Risk score with reasoning]
### Fund Flows
[Recent deposit/withdrawal trends]
### Recommendation
[Clear verdict: invest / avoid / monitor — with 1-sentence justification]
*Vault operator strategy is not fully transparent — current positions give a partial picture only. Past performance does not guarantee future results.*
Important Caveats
- Past vault performance does not guarantee future results — state this clearly
- Always recommend verifying performance figures on the Paradex UI before depositing —
include this in full analysis and any deposit recommendation ("Verify these figures on the Paradex vault page before depositing — data freshness and rounding can differ.")
- Vault token price can decline — depositors can lose money
- Withdrawal lockup periods apply — mention the vault's specific lockup
- This is analysis and screening, not investment advice
- Vault operator strategies are not fully transparent — always state this in every response. Positions show current holdings but do not reveal the full strategy or risk model.
- Small TVL vaults may have inflated ROI percentages from small base effects — always flag any vault with small TVL or limited drawdown history using the exact language: "insufficient drawdown history — ratio unreliable"
See [scoring.md](references/scoring.md) for detailed risk scoring methodology and JMESPath query cookbook.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: tradeparadex
- Source: tradeparadex/paradex-skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.