From 107ef83e08a20cd63a6f145edab6a5f83952c533 Mon Sep 17 00:00:00 2001 From: Dmitri Zaitsev Date: Fri, 16 May 2025 21:55:54 +0700 Subject: [PATCH] Create productContext.md --- cline_docs/productContext.md | 58 ++++++++++++++++++++++++++++++++++++ 1 file changed, 58 insertions(+) create mode 100644 cline_docs/productContext.md diff --git a/cline_docs/productContext.md b/cline_docs/productContext.md new file mode 100644 index 0000000..ffb5651 --- /dev/null +++ b/cline_docs/productContext.md @@ -0,0 +1,58 @@ +# productContext.md + +## What Is This Project? + +`aligna` is a meta-framework for AI-native software collaboration. + +It aims to reduce the need for repetitive prompting and manual instruction by embedding contextual knowledge directly within the project structure. + +At the same time, it introduces the concept of **external reviewers**—humans or AI agents that provide oversight, validation, and feedback to ensure quality, guide decision-making, and constrain ungrounded behaviors. + +## Why Does This Matter? + +As AI systems become more capable, we face a fundamental challenge: +> How can we **control the quality** of autonomous AI activity within complex systems? + +`aligna` addresses this by: +- Providing assistants with structured memory and navigation +- Delegating dynamic supervision to **reviewer agents** who can evaluate decisions, correct drift, and ensure adherence to purpose + +## Who Is This For? + +- Developers integrating AI into real software ecosystems +- Prompt engineers designing multi-agent workflows +- Reviewers (human or AI) assessing assistant outputs for accuracy, safety, or coherence +- Researchers exploring AI alignment through system-level design + +## Design Principles + +- **Static context guides active agents.** +- **Reviewers guide the agents.** +- **Bridge adapters connect agents to context.** +- **All logic is declarative and observable.** + +## Key File Anchors + +- `README.md`: Universal entrypoint +- `cline_docs/activeContext.md`: Current task state (bridge for Cline) +- `cline_docs/productContext.md`: Project identity and review philosophy + +## Long-Term Goal + +To enable reliable, explainable, tool-agnostic AI workflows grounded in: +- Declarative architecture +- Embedded context +- Independent review +- Promptless cooperation + +## Reviewer Layer (Emerging) + +This project invites the integration of **reviewer agents**—entities that: +- Evaluate outputs +- Guide assistants back toward goals +- Record decisions and interventions for traceability + +This layer will be prototyped via: +- Explicit reviewer prompts +- Human-in-the-loop validation +- Feedback-driven context updates