validate

Validate investigative claims against seven citation formats and contradictions.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/theLightArchitect/cappy-toolkit --skill validate-thelightarchitect
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: validate
Source: https://github.com/theLightArchitect/cappy-toolkit/tree/main/skills/validate
Command: npx skills add https://github.com/theLightArchitect/cappy-toolkit --skill validate-thelightarchitect

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Investigative outputs often contain uncited or improperly cited factual claims that can lead to incorrect triage, unnecessary escalations, or missed known issues. This Skill enforces citation completeness and correctness after Phase 3 evidence extraction so only verifiable claims proceed to Phase 4.

Core Features & Use Cases

  • Seven citation format validators: file:line, HAR entries, timestamps, pattern IDs, JIRA references, configuration keys, and environment metadata.
  • Contradiction detection: temporal, causal, factual, and evidence contradictions are identified and surfaced.
  • Verification rate & gating: computes a verification_rate and returns PASSED if ≥90% and no contradictions, otherwise returns BLOCKED with failed_claims and recovery options.
  • Hook integration: ClaimValidatorHook specification for structured agent invocation and response handling.
  • Use Case: Run after Phase 3 on an XSOAR/XSIAM investigation to ensure every claim is traceable to evidence before drafting customer responses or engineering escalations.

Quick Start

Validate the Phase 3 analysis for case SF-03845933 and return PASSED or BLOCKED with detailed failed claims, contradictions, and recovery options.

Frequently Asked Questions about validate

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I validate citations for investigative claims extracted from HAR files and logs?

Citation validation for investigative claims parses seven formats including HAR entries, file:line, timestamps, pattern IDs, JIRA references, configuration keys, and environment metadata. It detects temporal, causal, factual, and evidence contradictions before computing a verification rate.

What is the verification rate threshold for blocking uncited claims in an investigation?

The verification rate threshold for blocking uncited claims requires at least 90% verification and zero contradictions. If the rate falls below 90% or contradictions are detected, the validation returns BLOCKED with failed claims and recovery recommendations.

How do I detect contradictions in evidence extracted during XSOAR or XSIAM investigations?

Contradiction detection in extracted evidence identifies temporal, causal, factual, and evidence contradictions across parsed logs, HAR files, configuration data, JIRA references, and environment metadata to prevent incorrect triage or unnecessary escalations.

Does claim validation work with JIRA references and environment metadata from Phase 3 evidence extraction?

Claim validation supports JIRA references and environment metadata as two of seven parsed citation formats. It is designed to run after Phase 3 evidence extraction on XSOAR/XSIAM investigations to ensure every claim is traceable before drafting customer responses or engineering escalations.

Why does citation validation block investigative outputs with uncited factual claims?

Citation validation blocks investigative outputs with uncited factual claims to prevent incorrect triage, unnecessary escalations, or missed known issues. It enforces citation completeness and correctness so only verifiable claims proceed to Phase 4.