How does this skill prevent AI hallucinations?

It uses a mandatory 100% citation rule and verifies every claim through Ed25519 cryptographic signatures and citation chain integrity, ensuring the AI only reports facts it can prove.

What is GOAP in this context?

Goal-Oriented Action Planning allows the AI to dynamically plan its research steps, choosing the most cost-effective and trustworthy path to fill information gaps based on defined goals.

Can I customize the trusted sources?

Yes, the system includes a configurable Trusted Issuers Whitelist for academic, news, government, and financial domains, allowing you to restrict research to specific high-authority platforms.

What are the verification thresholds?

Users can set research modes from 'development' (0.75 confidence) up to 'paranoid' (0.99 confidence) depending on the required level of rigor for the specific task.

Verified Research & GOAP Verification

Name: Verified Research & GOAP Verification
Author: dzhechko

bydzhechko

0•

セキュリティとテスト

Conducts high-stakes research using Goal-Oriented Action Planning and Ed25519 cryptographic signatures to eliminate hallucinations and verify source authenticity.

This skill integrates Advanced Goal-Oriented Action Planning (GOAP) with Ed25519 cryptographic verification to provide a robust research framework. It is designed for contexts where factual accuracy is non-negotiable—such as legal, medical, or competitive intelligence—by enforcing a 100% citation rule and building verifiable signature chains for every claim. By combining the flexibility of AI research with the rigor of cryptographic proofs, it ensures that every piece of information is traced back to a trusted, signed source, significantly reducing the risk of AI-generated hallucinations and providing an auditable ledger of evidence.

主な機能

01Goal-Oriented Action Planning (GOAP) for adaptive research strategies

02Detailed verification ledgers and signed research reports

03Ed25519 cryptographic verification for source and claim authenticity

040 GitHub stars

05Automated building and verification of citation chains

06Configurable trust thresholds from developmental to paranoid modes

ユースケース

01Anti-hallucination protection for technical documentation and fact-checking

02High-stakes competitive intelligence with trusted issuer whitelisting

03Legal or medical research requiring verifiable evidence and audit trails

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add dzhechko/2026-jan-pu-intercom-clone-01 goap-research-ed25519

For use in Claude.ai and ChatGPT

主な機能

01Goal-Oriented Action Planning (GOAP) for adaptive research strategies

02Detailed verification ledgers and signed research reports

03Ed25519 cryptographic verification for source and claim authenticity

040 GitHub stars

05Automated building and verification of citation chains

06Configurable trust thresholds from developmental to paranoid modes

ユースケース

01Anti-hallucination protection for technical documentation and fact-checking

02High-stakes competitive intelligence with trusted issuer whitelisting

03Legal or medical research requiring verifiable evidence and audit trails

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add dzhechko/2026-jan-pu-intercom-clone-01 goap-research-ed25519

For use in Claude.ai and ChatGPT