What problem does it solve?
Main AI agents might provide filtered or summarized views of conversations, which can obscure critical details needed for thorough audits, debugging, or in-depth analysis of agent behavior and tool interactions. Relying on these summaries can lead to incomplete understanding or missed issues. This Skill provides direct, unfiltered access to the source.
Core Features & Use Cases
- Direct Log Access: Provides unfiltered access to raw conversation history files (
.jsonl), ensuring complete data for any analysis without agent interpretation.
- Structured Parsing Guidance: Offers
jq commands and patterns to efficiently extract specific event types, messages, tool calls, and working directory contexts from the raw JSON Lines format.
- Audit & Debugging Support: Essential for verifying user approval checkpoints, investigating tool errors, and understanding agent invocation sequences with unbiased, complete data.
- Use Case: An audit skill needs to verify if a user explicitly approved a critical step. Instead of relying on the main agent's summary, this Skill provides the raw conversation log, allowing the audit skill to independently search for specific user approval messages, ensuring unbiased verification and saving manual log review time.
Quick Start
Read the raw conversation history for the current session and extract all user messages.
