Tool Call Result Filtering And Output Redaction

via “intelligent-output-filtering-with-language-aware-parsing”

CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies

Unique: Uses a pluggable OutputParser framework with domain-specific filtering rules per command type (git, npm, python, etc.) rather than generic regex-based truncation. Preserves semantic information through language-aware parsing that understands tool output structure, enabling 60-90% reduction while maintaining LLM usability.

vs others: More sophisticated than simple output truncation or generic filtering — RTK's parser framework understands command semantics, achieving higher compression ratios while preserving critical debugging information that generic solutions would lose.

Core proxy engine for Cordon for MCP — the security gateway for MCP tool calls

Unique: Provides MCP-level output redaction that works across all tools without requiring per-tool implementation, enabling centralized data loss prevention and privacy enforcement

vs others: Redacts sensitive data at the protocol level after tool execution, whereas per-tool redaction requires implementing DLP in each tool and may allow sensitive data to leak through audit logs or monitoring

via “tool-call result inspection and output filtering”

The security gateway for AI agents — firewall, auditor, and remote control for MCP tool calls

Unique: Operates on tool results at the MCP protocol level, filtering before the agent receives data; supports both pattern-based detection (regex, data types) and custom validators for domain-specific sensitive data

vs others: More effective than agent-level filtering because it catches exfiltration attempts before the agent can log or process data; more transparent than application-level redaction because it operates at the gateway