Human Approval Gate Manager AIコネクタ
Manages deterministic human-in-the-loop approval gates for agent actions.
表示Memory Context Window Pruner AIコネクタ
Manage LLM conversation history by applying deterministic pruning strategies to prevent context window overflow.
表示Output Format Contract Checker AIコネクタ
Enforces strict data integrity and schema conformity between LLM pipeline stages.
表示Parallel Task Fan-out Coordinator AIコネクタ
Synchronize parallel agentic workflows with deterministic barrier strategies.
表示Tool Availability and Dependency Checker AIコネクタ
Audits tool environments for availability, version compatibility, and recursive dependency chains.
表示API Key Rotation Manager AIコネクタ
Deterministic API key rotation to prevent rate limits and exhaustion.
表示Deterministic Timeout Enforcer AIコネクタ
Prevents agent hang-ups by enforcing strict execution time boundaries and token budgets.
表示Dynamic Context Window Packer AIコネクタ
Optimally packs retrieved documents into LLM context windows using a greedy knapsack algorithm.
表示Graph State Merge Conflict Resolver AIコネクタ
Resolves deterministic merge conflicts for parallel graph node state updates.
表示Retrieval Relevance Scorer AIコネクタ
A deterministic scoring engine that filters retrieved documents using Jaccard, TF-IDF, and coverage metrics.
表示Streaming Chunk Assembler AIコネクタ
Reassembles fragmented LLM response chunks into complete, verified text and JSON.
表示Input Sanitizer & Escape Handler AIコネクタ
Protects AI agents from shell injection, path traversal, and payload vulnerabilities.
表示Truncation Detector & Graceful Terminator AIコネクタ
Detects and repairs truncated LLM outputs to restore structural integrity.
表示Agent Description & Tool Alignment Checker AIコネクタ
Validates alignment between agent descriptions and assigned tools to prevent multi-agent framework errors.
表示Constraint Satisfaction Verifier AIコネクタ
A deterministic validator for LLM outputs to enforce strict business and structural constraints.
表示Global Retry Budget Tracker AIコネクタ
Prevents unbounded retry loops by enforcing a strict credit-based budget across all agents.
表示json-state-diff-generator AIコネクタ
High-precision JSON state comparison for debugging agentic workflows.
表示Markdown Code Block Extractor AIコネクタ
Extracts exact code blocks from LLM markdown responses.
表示MinHash Text Deduplicator AIコネクタ
Detect near-duplicate texts using MinHash signatures and Jaccard similarity.
表示Python Code Safety Linter AIコネクタ
Validates Python code strings using AST analysis to prevent security vulnerabilities.
表示URL SSRF Prevention Validator AIコネクタ
Prevents SSRF attacks by validating URL schemes and blocking internal IP ranges.
表示XML to Dict Parser AIコネクタ
Converts XML strings into deterministic dictionary structures for stable LLM outputs.
表示Conversation Turn Fairness Enforcer AIコネクタ
Prevents agent domination in multi-agent conversations by monitoring turn frequency.
表示Document Metadata Flattener AIコネクタ
Flattens nested JSON metadata into dot-notation keys for vector store compatibility.
表示LLM API Cost Estimator AIコネクタ
Calculate exact API costs and track session budgets for LLM workflows.
表示Resource Limit Calculator AIコネクタ
Translates high-level resource constraints into exact OS-level RLIMIT constants.
表示Routing Logic Evaluator AIコネクタ
A decision engine for deterministic conditional routing in multi-agent workflows.
表示Task IO Mapping Validator AIコネクタ
Validates data continuity and key mapping between sequential tasks in AI agent pipelines.
表示Template Injection Sanitizer AIコネクタ
Detects and neutralizes prompt injection attacks using template syntax like Jinja2 and Python f-strings.
表示Tool Namespace Resolver and Fuzzy Matcher AIコネクタ
Resolves tool name hallucinations and misspellings using hierarchical matching logic.
表示Tool Output Truncator and Size Limiter AIコネクタ
Safely shrinks tool outputs to fit byte limits without breaking multi-byte characters.
表示Agent Config Drift Detector AIコネクタ
Detects unauthorized changes to agent configurations by comparing SHA-256 hashes.
表示
