The Agent Skills Directory

Diagnostic Command Execution: The skill recommends using shell commands like jq or grep, or writing small Python scripts to parse transcript.jsonl and conversation.jsonl. This is an efficient way to extract specific events from large log files without exceeding context window limits.
Processing of Agent Transcripts: The skill ingests data from previous agent stages to identify successes and failures. This analysis of historical data is the core function of the reflection stage.
Ingestion points: Conversation and transcript logs (transcript.jsonl, conversation.jsonl) referenced in the SKILL.md file.
Boundary markers: The instructions do not define specific delimiters for separating processed data from agent instructions.
Capability inventory: The skill utilizes bash and Python script execution capabilities for data filtering (SKILL.md).
Sanitization: No explicit sanitization or filtering of the ingested log content is mentioned in the instructions.

mantis-reflect