Log post-task performance metrics and session data for prompt decisions.
npx skills add https://github.com/hjung3113/vocpage --skill opt-prompt-eval
Please help me install this Agent Skill. Skill: opt-prompt-eval Source: https://github.com/hjung3113/vocpage/tree/main/.claude/skills/opt-prompt-eval Command: npx skills add https://github.com/hjung3113/vocpage --skill opt-prompt-eval
This Skill enables precise recording of post-task retrospective data for normalized prompts, maintaining accountability and improving prompt workflows.
Invoke /opt-prompt-eval <decision_id> after prompt completion to log retro data for analysis.
High-intent search queries and answers about installing and using this skill.
To log prompt evaluation metrics after a task completes, invoke the retro logging command with a specific decision ID to automatically write a JSONL row capturing session data and performance insights.
Ensuring traceability for prompt decisions involves linking retrospective log entries to specific decision IDs and session contexts, providing structured accountability for post-task performance analysis.
Recording post-task performance data like latency and token usage is handled by automated Python scripts and shell helpers that write structured JSONL logs for continuous prompt optimization.
Yes, you need a specific decision ID to start logging retrospective prompt data, as passing this identifier triggers the automated capture of session metrics and ensures data traceability.
You need to use JSONL logs for prompt evaluation when your team requires structured retrospective insights and data integrity for analyzing post-task performance metrics and refining prompt workflows.
Limitations of using shell helpers for prompt evaluation logging include dependencies on specific script components like append.sh and session-stats.sh to maintain data consistency and execute the automated logging process correctly.
yzhao062Ingest NSF solicitation text into a normalized requirements baseline.
dakshrawat298-gifAnalyze session JSONL logs with jq and rg to extract costs, token usage, and tool interactions.
tomgunResume interrupted development sessions and present project state with next steps.
Ankh-StudioProcess multiple artifacts concurrently to generate comparative analysis and trend reports.
