What are Agentic KPIs?

Agentic KPIs are specialized metrics used to measure the efficiency and autonomy of AI agents, focusing on success rates, the number of retries (Presence), and the complexity of tasks handled.

How is the 'Presence' metric calculated?

Presence represents the average number of attempts or workflow restarts required to complete a task. A target Presence of 1.0 indicates perfect first-time success.

Why should I track streaks?

Tracking success streaks helps determine the reliability of the agent. A consistent streak of successes (e.g., 5 or more) is often a key indicator that the workflow is ready for autonomous operation.

What does ZTE stand for in this context?

ZTE stands for Zero-Touch Engineering, which refers to a level of maturity where an AI agent can complete a coding task successfully from start to finish without human intervention.

Where is the KPI data stored?

The skill typically logs and updates metrics in a local file, such as app_docs/agentic_kpis.md, allowing the data to persist and be tracked within your version control system.

Agentic KPI Tracking

Name: Agentic KPI Tracking
Author: melodic-software

bymelodic-software

•

분석 및 모니터링

Measures and monitors autonomous coding performance metrics to track progress toward Zero-Touch Engineering (ZTE).

The Agentic KPI Tracking skill provides a structured framework for measuring the effectiveness of AI-driven development workflows. By tracking specific metrics such as success streaks, attempt counts (Presence), and change magnitudes, it allows developers to quantify the reliability and efficiency of their agentic processes. This skill is essential for teams looking to transition from human-in-the-loop interactions to autonomous operation, providing the data-driven insights needed to verify Zero-Touch Engineering readiness.

주요 기능

01Automated tracking of success streaks and attempt counts per coding task

02Standardized markdown reporting for KPI persistence and documentation

03Calculation of Presence metrics to measure workflow efficiency

04ZTE readiness assessment based on historical performance trends

05Detailed analysis of Plan and Diff sizes to track throughput and complexity

061 GitHub stars

사용 사례

01Benchmarking the reliability of AI-driven coding agents over time

02Identifying task categories where agents require the most human intervention

03Building data-backed confidence for moving to autonomous execution (ZTE)

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add melodic-software/claude-code-plugins agentic-kpi-tracking

For use in Claude.ai and ChatGPT

주요 기능

01Automated tracking of success streaks and attempt counts per coding task

02Standardized markdown reporting for KPI persistence and documentation

03Calculation of Presence metrics to measure workflow efficiency

04ZTE readiness assessment based on historical performance trends

05Detailed analysis of Plan and Diff sizes to track throughput and complexity

061 GitHub stars

사용 사례

01Benchmarking the reliability of AI-driven coding agents over time

02Identifying task categories where agents require the most human intervention

03Building data-backed confidence for moving to autonomous execution (ZTE)

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add melodic-software/claude-code-plugins agentic-kpi-tracking

For use in Claude.ai and ChatGPT