experience-curator — Claude Code subagent: what it does
experience-curator — a Claude subagent that screens finished task runs and keeps only verified lessons in ReasoningBank memory.
no license ★ 850 updated 2026-10-26
.claude/agents/experience-curator.mdWhat experience-curator does
The subagent reads a task history and decides which runs earn a place in ReasoningBank memory. Each run is scored on clarity, reliability, actionability, generalizability and novelty, then weighted into one number. At 0.7 or above the entry is accepted, from 0.5 it goes to review, and anything lower is dropped.
Two hooks in the frontmatter do the fetching. Before the task the agent pulls earlier runs with npx agentic-flow@latest reasoningbank retrieve "$TASK" --with-outcomes; afterwards it distills the result using --quality-filter high. Screening throws out abandoned tasks, runs with no clear verdict and duplicates of patterns already stored. Accepted findings follow a fixed shape: insight, context, evidence, limits and the steps to apply them.
The 0.7 cut is strict. Memory stays clean, yet mid-grade lessons never make it in, and getting them through means editing the weights in the agent file. Both hooks call agentic-flow through npx, so a broken npm setup leaves the subagent as plain text.
What it helps with
- filter out lessons that duplicate stored memory
- salvage a usable takeaway from a failed run
- judge whether a finding generalizes
- choose what from a task history belongs in memory
- tell a one-off win from a repeatable trick
Example prompts for the agent
- Run
experience-curatorover the last task history and keep only entries above the cut. - Pull a lesson worth storing out of the failed run and add it to ReasoningBank.
- Check the new entries for duplicates against what memory already holds.
How to install experience-curator
Manually: put the files in place
.claude/agents/experience-curator.mdFor all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention
Manually: put the files in place
.codex/agents/experience-curator.tomlCodex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/
Manually: put the files in place
.cursor/agents/experience-curator.mdCursor also reads .claude/agents/
Step by step, with commands for each platform: Claude Code subagents: how to create one, call it and keep context clean
Questions about experience-curator
How do I install experience-curator in Claude Code?
Copy the subagent file into .claude/agents/experience-curator.md inside your project. Claude loads subagents from that folder, so nothing else is required.
What does experience-curator do?
It reviews finished task runs and decides which lessons earn a spot in ReasoningBank memory, dropping weak, duplicate or abandoned ones.
What do the 0.7 and 0.5 scores mean?
0.7 or higher accepts the entry, 0.5 to 0.7 sends it to review, and anything below is rejected. Reliability carries the largest weight in the total.