# Context Provenance Context provenance is the practice of tracking where each piece of AI context came from, when it was added, by whom, and how trustworthy it is. It answers the question: "Why is this in the context, and should I trust it?" ## Why provenance matters AI models treat all context as equally trustworthy. They have no built-in mechanism to distinguish between a carefully reviewed CLAUDE.md rule and a scraped web page that happened to land in a RAG result. Context provenance adds the metadata layer that models lack. Without provenance: - [[Context Poisoning]] is harder to detect (you can't trace where bad information entered) - [[AI Context Rot]] is harder to fix (you don't know when entries were last validated) - [[AI Context Governance]] is impossible to enforce (you can't audit what you can't trace) - Debugging AI behavior is guesswork (you can't reconstruct what the model saw) ## Provenance dimensions | Dimension | Question | Example | |---|---|---| | **Source** | Where did this come from? | Human-authored, RAG retrieval, tool output, AI-generated memory | | **Author** | Who created or approved it? | User, team lead, automated pipeline | | **Timestamp** | When was it added/updated? | Created 2026-01-15, last reviewed 2026-03-01 | | **Trust level** | How reliable is this source? | Verified (human-reviewed) vs unverified (auto-retrieved) | | **Scope** | Who is this for? | Enterprise, team, personal, task-specific | ## Practical implementation In [[Context-as-Code]] systems, provenance comes naturally through version control: git blame shows who changed what and when. For dynamic context (RAG results, tool outputs, AI memories), provenance requires explicit metadata. The [[Context Lifecycle]] review phase is where provenance becomes actionable: during review, you can check whether context entries still have valid provenance or have become orphaned. ## References - ## Related - [[Context Engineering]] - [[Context Poisoning]] - [[AI Context Rot]] - [[AI Context Governance]] - [[Context-as-Code]] - [[Context Lifecycle]] - [[Context Isolation]] - [[Context Hygiene]] - [[Personal Context Management (PCM)]] - [[Enterprise Context Management (ECM)]]