To maintain a professional, scalable, and traceable audit trail of AI contributions, Nox uses a decentralized logging system. Every agent MUST follow this format to ensure accountability and clarity for both humans and other AI agents.
Logs are no longer stored in a single root file. They are organized by agent identity:
AGENTS_LOGS/<Agent_Name>/AGENTS_LOGS.md
Each log entry MUST use the following Markdown structure:
# [YYYY-MM-DD] Log Entry XX: <Brief Title>
**Agent Identity:** <Agent Name> (e.g., Gemini CLI, Jules, GPT-5.3-Codex)
### 1. User Instructions (Directives)
* **Request:** <The exact or summarized request from the user>
* **Constraints:** <Any specific constraints or rules mentioned by the user>
### 2. Technical Implementation Details
* **Architecture Changes:** <What was changed at a system level>
* **Files Modified:** <List of key files>
* **Logic Forensics:** <Reasoning behind specific code choices>
### 3. Error Recovery & Course Corrections (Self-Audit)
* **Mistakes Identified:** <Any errors made during execution, e.g., git mistakes, bugs introduced>
* **Remediation:** <How the error was fixed and what was learned>
### 4. Final Verification
* **Tests Run:** <List of tests executed>
* **Success Criteria:** <Evidence of 100% completion>
**Status:** <e.g., Complete, Hardened, Production Ready>
**Author:** <Signature>- Atomic Logs: Create a new entry for every significant task or "chapter" of work.
- Honesty in Recovery: If a mistake is made (e.g., a broken build, a force-push, a logic bug), it MUST be documented in Section 3.
- User Instruction Tracking: Always quote or summarize the user's intent to provide context for the changes.
- No Deletion: Never delete previous logs; only append new ones or fix typos if absolutely necessary.