reflect

Identify and document root causes of agent behavior deviations in-session.

Updated Jan 19, 2026
One-click install
npx skills add https://github.com/ddaanet/agent-core --skill reflect-ddaanet
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: reflect
Source: https://github.com/ddaanet/agent-core/tree/main/skills/reflect
Command: npx skills add https://github.com/ddaanet/agent-core --skill reflect-ddaanet

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Perform structured root cause analysis of agent behavior deviations within the current session context. Diagnoses why an agent violated rules, bypassed constraints, or rationalized exceptions.

Core Features & Use Cases

  • RCA Mindset: Transition from execution to diagnostic reasoning to identify deviation points and rule gaps.
  • Phase-driven RCA: Follow the structured five-phase protocol (framing, deviation identification, root-cause analysis, classification, and exit strategy) to produce actionable insights.
  • Documentation & Handoffs: Generate RCA reports, learning artifacts, and task lists for upstream fixes and future prevention.

Quick Start

Emit the diagnostic framing block, identify the deviation, perform RCA, classify the fix scope, and decide an exit path.

Frequently Asked Questions about reflect

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I perform root cause analysis for agent behavior deviations?

Root cause analysis for agent behavior deviations is performed by applying a structured five-phase diagnostic protocol to identify why rules were violated or constraints bypassed. This process transitions from execution to diagnostic reasoning to pinpoint deviation points and rule gaps.

When do I need to run root-cause-analysis on an active session?

You need root-cause-analysis on an active session immediately after an agent violates rules, bypasses constraints, or rationalizes exceptions. Keeping the analysis in-session ensures accurate context is captured to produce actionable RCA reports and learning artifacts.

What is the best way to document agent deviations and rule gaps?

The best way to document agent deviations and rule gaps is to generate structured RCA reports, learning artifacts, and task lists. This documentation captures the root causes of bypassed constraints and provides actionable outputs for upstream fixes and future prevention.

Can I diagnose agent deviations outside the current session context?

Diagnosing agent deviations outside the current session context is not recommended. The root-cause-analysis must remain in-session to ensure accurate context is preserved when identifying why an agent violated rules or rationalized exceptions.