「detector」の検索結果

Detects prompt injection attacks targeting LLM-based applications using a multi-layered defense combining regex pattern matching for known attack signatures, heuristic scoring for structural anomalies, and transformer-based classification with DeBERTa models. The detector analyzes user inputs before they reach the LLM, flagging direct injections (system prompt overrides, role-play escapes, instruction hijacking) and indirect injections (encoded payloads, multi-language obfuscation, delimiter-based escapes). Based on the OWASP LLM Top 10 (LLM01:2025 Prompt Injection) and Simon Willison's prompt injection taxonomy. Activates for requests involving prompt injection detection, LLM input sanitization, AI security scanning, or prompt attack classification.

日本語の概要は準備中です。原文の説明を表示しています。

Weekly churn risk detection across active client accounts. Scans for signals like decreased engagement, missed meetings, delayed payments, and competitor mentions. Scores risk 1-10 and outputs prioritized alert list.

日本語の概要は準備中です。原文の説明を表示しています。

Validates agent outputs against expected schemas and quality criteria. Reports structural and content checks within the declared evidence scope. Activate on 'validate output', 'output validation', 'schema validation', 'check output', 'output quality'. NOT for confidence scoring (use dag-confidence-scorer) or hallucination detection (use dag-hallucination-detector).

日本語の概要は準備中です。原文の説明を表示しています。

Synthesizes evidence-linked feedback from validation results, confidence evidence, and iteration triggers. Proposes bounded revision guidance for a stated acceptance failure. Activate on 'synthesize feedback', 'improvement suggestions', 'actionable feedback', 'iteration guidance', 'feedback generation'. NOT for iteration detection (use dag-iteration-detector) or convergence tracking (use dag-convergence-monitor).

日本語の概要は準備中です。原文の説明を表示しています。

Performs claim-evidence verification for agent outputs, recording supported, contradicted, or insufficient-evidence results alongside source accessibility. Activate on 'detect hallucination', 'fact check', 'verify claims', 'check accuracy', 'find fabrications'. NOT for validation (use dag-output-validator) or confidence scoring (use dag-confidence-scorer).

日本語の概要は準備中です。原文の説明を表示しています。

Identifies when task outputs require iteration based on quality signals, unmet requirements, or explicit feedback. Triggers appropriate re-execution strategies. Activate on 'needs iteration', 'retry needed', 'not good enough', 'try again', 'refine output'. NOT for feedback generation (use dag-feedback-synthesizer) or convergence tracking (use dag-convergence-monitor).

日本語の概要は準備中です。原文の説明を表示しています。

Analyze repeated non-engagement, deflection, motte-and-bailey shifts, tone policing, sealioning, and related manipulative discourse patterns. Use when the user suspects they are being rhetorically managed rather than answered, or when they need a careful pattern read grounded in quotes and alternatives. NOT for psychiatric diagnosis, legal judgment, or help deploying bad-faith tactics.

日本語の概要は準備中です。原文の説明を表示しています。