Loading ToolBench
Solana Token & NFT Safety Oracle — 12-layer rug pull, honeypot, LP lock, deployer recon, and NFT analysis for AI agents via MCP
SicariusGuard defines 6 tools with complete input schemas and detailed descriptions. All tools follow verb_noun naming (check_*, recon_*, full_*) and have descriptions 150-400 chars, well within the 10-1024 baseline. Schemas are properly typed with required fields. However, output schemas are NOT documented, LLMs cannot see what fields to expect from responses. Parameter descriptions are good but lack explicit constraints (e.g., no enum for verdict values, no format spec for mint addresses). Error handling and recovery guidance are absent from tool descriptions. No tool annotations (readOnlyHint, idempotentHint) despite all being read-only operations.
Analyze the distribution of token holders for a Solana SPL token. Detects if supply is dangerously concentrated in a few wallets, which is a common rug pull setup. Flags as concentrated if: - Top 1 holder owns >50% of supply - Top 5 holders own >80% of supply - Top 10 holders own >90% of supply Returns the top 10 holders with their addresses, amounts, and percentage of total supply.
Check if a Solana token is a honeypot by simulating a sell order through Jupiter. If Jupiter cannot find a route to sell the token back to SOL, it is likely a honeypot — meaning you can buy but cannot sell. This is a quick, zero-cost check (quote only, no swap executed). Use this for a fast preliminary filter before doing a full safety analysis.
Perform a safety analysis on a Solana NFT using on-chain DAS metadata and Magic Eden marketplace data. Detects counterfeit collections, unverified creators, wash-traded NFTs, burnt/frozen assets, excessive royalties, and pricing anomalies. Analyzes 7 risk signals: 1. Collection verification — is the collection group verified on-chain? 2. Creator verification — are the NFT creators verified? 3. Metadata mutability — can the creator change the NFT's image/name after sale? 4. Marketplace presence — is the NFT listed on Magic Eden? At what price? 5. Collection health — floor price, volume, listing count 6. Asset status — compressed, burnt, frozen 7. Pricing anomalies — listed far above/below floor price Returns a JSON object with riskScore (0-100), verdict (SAFE/CAUTION/HIGH_RISK/CRITICAL), collection data, floor price, and per-signal flags. This is a read-only operation with no on-chain side effects. Use this for NFT safety checks; use check_token_safety for fungible SPL tokens instead.
Output schemas not documented. LLMs cannot see what fields responses contain (e.g., does check_token_safety return 'riskScore' or 'risk_score'? Does recon_deployer return 'recidivismScore' or 'recidivism_score'?). This forces agents to guess field names and breaks downstream tool chaining.
Parameter constraints not formalized. 'mint' is described as 'base58 encoded' but no regex pattern or format constraint. 'verdict' values (SAFE/CAUTION/HIGH_RISK/CRITICAL) are mentioned in descriptions but not declared as enums. LLMs cannot validate inputs before calling.
No error handling or recovery guidance in tool descriptions. If a mint address is invalid or a token does not exist, what should the LLM do? Descriptions do not explain failure modes or next steps.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 62 | 2026-07-28+ | v2 |
Analyze a Solana SPL token for rug pull, honeypot, and safety risks. Call this BEFORE executing any swap or buy transaction to protect against scams. Performs 5 checks: 1. Mint authority — can the deployer print infinite tokens? 2. Freeze authority — can the deployer freeze your wallet? 3. Token-2022 extensions — PermanentDelegate, TransferHook, ConfidentialTransfers traps 4. Honeypot simulation — can you actually sell this token back to SOL? 5. Holder concentration — is supply concentrated in a few wallets? Returns a combined risk score (0-100), verdict (SAFE/CAUTION/HIGH_RISK/CRITICAL), and per-check details.
Perform the most comprehensive safety analysis available on a Solana token by combining on-chain byte-level inspection with Birdeye market intelligence. This is the premium check — use it for high-value trades or when you need maximum confidence. Includes everything from check_token_safety PLUS: - Birdeye token overview (price, volume, liquidity, market cap) - Birdeye security flags (holder distribution, mutable metadata) - Birdeye trade data (wash trading detection, bot activity) - Market risk scoring with weighted final verdict Returns a dual risk score: - On-chain risk score (0-100) - Market risk score (0-100) - Weighted final score (70% on-chain, 30% market)
Perform a deep background investigation on a Solana wallet that deployed or controls a token. Returns a complete dossier including their token creation history, portfolio health, wallet age, funding genesis, and a Recidivism Score (0-100). Use this BEFORE trusting any new token — if the deployer has a history of abandoned or rugged tokens, the current token is likely another scam. Analysis pipeline: 1. Portfolio enumeration — finds all fungible tokens the deployer controls or has created 2. Health triage — classifies each token as alive, abandoned, or rugged 3. Wallet genesis — identifies who funded the deployer and when 4. Recidivism scoring — algorithmic risk assessment based on dead token ratio, deployment frequency, wallet age, and authority patterns Verdicts: CLEAN | SUSPICIOUS | SERIAL_DEPLOYER | LIKELY_SCAMMER This is a zero-cost check using DAS API and Enhanced Transactions (Helius Developer plan).
No tool annotations despite all tools being read-only. Adding readOnlyHint=true to all 6 tools would signal to agents that these are safe to call without side effects and can be retried freely.
Parameter 'amount' in check_honeypot lacks type and range constraints. Description says 'raw token amount' but does not specify: is it a string or number? What is the valid range? Default is 1000000 but no min/max documented.