evidence-verifier

Verifies artifact integrity, provenance chains, and redaction before graph promotion decisions.

10|3|Updated Aug 10, 2026
One-click install
npx skills add https://github.com/baiqigo/baiqi-redteam-lab --skill evidence-verifier-baiqigo
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: evidence-verifier
Source: https://github.com/baiqigo/baiqi-redteam-lab/tree/main/.agents/skills/evidence-verifier
Command: npx skills add https://github.com/baiqigo/baiqi-redteam-lab --skill evidence-verifier-baiqigo

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Security research pipelines risk promoting unverified claims into knowledge graphs. This Skill validates that every artifact and claim has matching hashes, complete provenance, and proper redaction before any graph edge or proof condition is promoted. ## Core Features & Use Cases - Artifact Integrity Verification: Resolves every artifact in the Case evidence boundary and confirms path, size, and SHA-256 match, rejecting mismatches. - Provenance Tracing: Traces each claim back to its producing task, run, revision, and observation so no claim remains implicit. - Redaction Enforcement: Rejects credential-bearing evidence or represents it only as a redacted classification. - Use Case: A supervisor assigns the evidence.verify capability during a red-team exercise; the verifier checks all collected artifacts and returns promotion decisions for each claimed graph edge in the worker-result schema. ## Quick Start Ask the agent to run the evidence-verifier skill on the current case evidence and return promotion decisions for all claimed graph edges.

Frequently Asked Questions about evidence-verifier

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

FAQPage Schema
How do I verify artifact integrity before promoting evidence to a knowledge graph?

Resolve every artifact inside the case evidence boundary and confirm its path, size, and SHA-256 hash match the recorded values. Artifacts that fail any check are rejected and excluded from promotion decisions.

What is evidence provenance tracing in security research pipelines?

Provenance tracing links each claim to its producing task, run, revision, and observation. A claim is only accepted when this chain is complete, ensuring no implicit or untraceable assertions enter the graph.

How does the verifier handle evidence containing credentials?

Credential-bearing evidence is either rejected outright or represented only by a redacted classification. This enforces redaction and reproducibility requirements before any promotion decision is returned.

What output format does the evidence verification step return?

It returns a result conforming to the worker-result-v1 JSON schema, echoing the skill ID evidence.verifier, version 1.0.0, and its digest through the assigned task anchors, with a promotion decision per claimed graph edge.

When should evidence verification run in a multi-agent workflow?

Run it when a supervisor assigns the evidence.verify capability, after artifacts are collected but before graph promotion. It acts as a gate so only verified, provenance-complete claims become graph edges or proof conditions.