# Techstack Correlation > Correlation, confidence scoring, and conflict resolution across all tech-stack signals. Cross-validates and assigns High/Medium/Low confidence per technology. Use when this capability is needed. - Skill: `tomevault-io/techstack-correlation` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add tomevault-io/techstack-correlation` - Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/techstack-correlation/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: tomevault-io (https://skillmd.com/u/tomevault-io) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/tomevault-io/techstack-correlation --- # Tech-Stack Correlation ## Scope The Phase 4 layer that consumes all detection outputs (frontend, backend, infra, security, OSINT) and produces the final vetted technology list: - **Cross-validate** signals from independent sources for corroboration - **Score confidence** per technology (multi-factor weighted scoring) - **Resolve conflicts** between signals (version mismatch, mutually exclusive frameworks, temporal inconsistencies, categorical ambiguity) - **Flag** unresolvable conflicts for manual review ## Signals (input) - Raw signals grouped by category: `http`, `dns`, `tls`, `javascript`, `html`, `repository`, `job`, `archive` - Inferred technologies from each detection domain (frontend / backend / infra / security) - Asset inventory (domains, subdomains, IP map) for URL-context resolution ## Inferences (output) - `correlated_technologies[]` — each with `sources`, `source_count`, `agreement_score`, `corroborating_signals` - `conflicts[]` — typed (version_mismatch / mutually_exclusive / temporal_inconsistency / categorical_ambiguity), with severity - `technologies_with_confidence[]` — confidence level (High / Medium / Low), score (0-1), reasoning breakdown - `confidence_summary` — counts, overall score, per-category breakdown, recommendations - `confidence_gaps[]` — technologies needing additional evidence with suggested actions ## Techniques See [reference/patterns.md](reference/patterns.md). ## When to use - Always — last analysis stage before report generation - Required to decide what makes the final tech-stack list and at what confidence - Pure computation; no external requests --- > Source: [transilienceai/communitytools](https://github.com/transilienceai/communitytools) — distributed by [TomeVault](https://tomevault.io).