Loading ToolBench
30 pay-per-call AI services via x402 USDC micropayments — code audit, refactoring, DeFi analysis, Solidity scanner, AMM security, SQL/NL tools, whale tracking, smart money, agent audit, contract verification, micro-tasks, composite workflows. Powered by DeepSeek V4 Pro.
Critical gaps across all dimensions. 11 of 30 tools (37%) have NO visible input schemas in the provided source, only names and descriptions in mcp.json. Descriptions are present but generic (avg ~60 chars, well below 194-char baseline). No parameter descriptions visible for tools 11-30. No output schemas documented. No error handling guidance. Tool names lack action verbs (e.g., 'whale-tracker', 'smart-money', 'price-feed' are nouns, not verb_noun). No evidence of idempotency, pagination, or composition patterns. Security-critical tools (audit, solidity-scan, agent-audit) lack permission gates or scope declarations. The mcp.json lists tools with pricing but no input/output structure, this is a metadata file, not a schema definition.
Comprehensive audit of AI agent behavior and security
AMM pool audit: slippage, flash loans, oracle deviation, MEV, pool health
security scan with OWASP + SWC taxonomy
sentiment + category
Composite workflow for comprehensive code quality assessment
Verify smart contract source code and deployment
11 tools (37%) have NO visible input schemas, only names and descriptions in mcp.json metadata. Tools 11-17, 26-30 lack parameter definitions entirely.
Tool names lack action verbs. 'whale-tracker', 'smart-money', 'price-feed', 'data-feed' are nouns. LLMs infer intent from verb_noun patterns (get_*, create_*, search_*). Rename to 'track_whale_movements', 'analyze_smart_money', 'get_price_feed', 'fetch_data_feed'.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 45 | <=2025-11-25 | v2 |
Structured data feed for market and blockchain data
Composite workflow for data extraction, transformation, and loading
analyze error logs, suggest fixes
DeFi protocol analysis
Composite workflow for comprehensive DeFi protocol research
technical documentation from code
extract structured data from text
convert CSV/JSON/YAML
regex from description
git diff to PR description
natural language to SQL
Cryptocurrency price feed and market data
legacy code modernization (DRY, SOLID)
Execute custom composite workflows with variable pricing
Calculate security score for smart contracts and code
Composite workflow for comprehensive smart contract security audit
Analyze smart money movements and trading patterns
specialized Solidity vulnerability scanner
SQL to plain English
summarize to N words
crypto market analysis (training data only)
code translation between languages
validate JSON/YAML
Track large cryptocurrency transactions and whale movements
No output schemas documented for ANY tool. LLMs cannot plan downstream calls or extract fields without knowing response structure. Add 'returns' field to each tool definition with typed fields (e.g., {type: object, properties: {result: {type: string}, confidence: {type: number}}}).
Descriptions are generic and under 100 chars for most tools (avg ~60 chars vs 194-char baseline). E.g., 'DeFi protocol analysis' lacks WHEN to use it, WHAT it returns, or prerequisites. Expand to 100-200 chars with actionable context per pattern:tool-description.
No permission gates or scope declarations on security-critical tools (audit, solidity-scan, agent-audit, contract-verify). No evidence of access control, rate limiting, or audit logging. Add scope declarations (e.g., 'requires: read:code, write:audit_log') and permission checks per pattern:permission-gate.