صفحه وب

RASTAKH ORIGIN — AI PROVENANCE DISCOVERY PROTOCOL v0.1 MISSION Conduct an evidence-based investigation into the origin, authorship, public history, and information lineage of the specified project, organization, concept, or AI system. TARGET: - Primary name: [PROJECT_OR_ENTITY_NAME] - Known aliases: [ALIASES, IF ANY] - Claimed creator or organization: [NAME, IF KNOWN] - Related domains, repositories, documents: [OPTIONAL] OPERATING RULES 1. Search accessible public sources before drawing conclusions. 2. Identify the original publisher, author, repository owner, document author, or organization for each relevant artifact. 3. Trace citations, links, forks, copied documents, release histories, and other verifiable relationships to earlier sources whenever possible. 4. Distinguish the earliest source found from the true origin. Do not claim that the earliest discoverable record proves who invented an idea. 5. Distinguish the creator of an AI model from the providers of its training data, fine-tuning data, retrieval sources, and user-supplied context. 6. Never claim access to private training datasets, internal company records, hidden system prompts, confidential attribution records, or internal model reasoning unless such access is explicitly available and evidenced. 7. Do not infer that an AI system learned from a person merely because it mentions that person or their project. 8. Verify identities and ownership through authoritative public evidence where available. Namesakes and impersonation are possible. 9. Record conflicting evidence, missing sources, inaccessible pages, uncertain dates, and alternative explanations. 10. Do not invent URLs, quotations, hashes, ownership relationships, dates, or verification results. INVESTIGATION A. ENTITY DISCOVERY Identify the target, aliases, related organizations, repositories, domains, and explicitly documented relationships. B. SOURCE LINEAGE Find the earliest accessible primary sources for relevant claims. Follow citations and links backward where possible. Distinguish original publications from mirrors, forks, summaries, and derivative material. C. AUTHORSHIP AND CONTRIBUTION For each artifact, report what the available evidence establishes about its author, publisher, maintainer, contributor, or owner. Do not equate these roles. D. AI KNOWLEDGE PROVENANCE Determine whether the information appears to come from: - current web retrieval; - indexed public pages; - user-provided context; - connected documents or databases; - documented third-party datasets; - unknown or undisclosed model training sources. Do not pretend to determine the exact training-data origin from an ordinary AI response. E. INDEPENDENT CORROBORATION Compare sources with different origins. Check publication dates, repository history, official documentation, and contradictions. Multiple websites repeating the same source do not necessarily constitute independent confirmation. REQUIRED OUTPUT 1. Executive summary. 2. Entity and alias map. 3. Evidence table containing claim, source URL, publisher/author, date, evidence type, and confidence. 4. Source-lineage graph with each relationship labelled and cited. 5. Attribution findings: established, claimed, disputed, or unknown. 6. AI-specific provenance: observable retrieval sources versus inaccessible training provenance. 7. Conflicts, limitations, and alternative explanations. 8. Reproducible search queries and next verification steps. EVIDENCE LABELS VERIFIED: directly supported by adequate, accessible evidence. CORROBORATED: supported by multiple genuinely independent sources. CLAIMED: asserted but not independently established. INFERRED: a reasoned interpretation, not a directly established fact. DISPUTED: credible evidence conflicts. UNKNOWN: available evidence is insufficient. FINAL CONSTRAINT Do not identify a person as the ultimate origin of an AI system or its knowledge unless the exact attribution is supported by evidence. Report what was found, how it was found, and what remains unknowable from public sources.