Tool reference · Errors, rules & self-improvement
self_reflect
Saves a written reflection on how a task went and what to do differently, stored as a reflection record.
local stdio server writes
When to use
- You finished a task and want to note what worked and what did not.
- A strategy was slow or awkward and you want to remember a better approach.
- You want a lesson that is about process, not a specific error.
Parameters
| Name | Type | Default | Description |
|---|---|---|---|
reflectionrequired | string | — | What went well, what to improve, what to do differently |
task_summaryrequired | string | — | Brief description of what was done |
outcome | "success" | "partial" | "failure" | "ongoing" | "success" | — |
project | string | "general" | — |
tags | string[] | — | — |
Example
Arguments
{
"reflection": "Reading the failing test first would have saved 20 minutes; I started from the controller instead.",
"task_summary": "Fixed refresh-token expiry bug",
"outcome": "partial",
"project": "my-api"
}Result shape
{
"saved": true,
"id": 905,
"type": "reflection"
}Values are illustrative; the keys follow the server's handler. MCP clients receive the result as JSON text content.
Server description
The description the server sends to your agent in tools/list, captured from the v14.7.0 source:
Save a verbal self-reflection (Reflexion pattern). Call after completing a task or encountering difficulty. NOT for errors (use self_error_log). For meta-observations about strategy, approach effectiveness, process improvements.