Skip to content
Get packs

Guide · Agent skill

How to use the VP Engineering Advisor agent skill (vpe-advisor)

Updated · written with AI from the project's own documentation and checked against our catalogue data

The vpe-advisor skill gives an AI agent a structured playbook for the operational side of running an engineering team: how fast code ships, where engineering hiring leaks candidates, how squads and tech-lead roles should be arranged, and how on-call and postmortem practices look. It is written for startup VPs of Engineering and founders who do not have one yet. After this guide you will know what an agent skill is, how to install vpe-advisor, how to ask for help in plain language, and what the skill will and will not do.

VP Engineering Advisoralirezarezvani/claude-skills · MIT

Quick start

  1. 1Download the ZIP. Download the vpe-advisor ZIP for c-level-advisor/skills/vpe-advisor from the skill's page.
  2. 2Unzip to a skills folder. Place the folder in ~/.claude/skills/ for personal use, or in a project's .claude/skills/ folder for one project.
  3. 3Or upload in the Claude apps. In the Claude apps, open Settings, then Capabilities, then Skills, and upload the skill.
  4. 4Ask in plain language. Describe the problem, such as slipping sprint velocity or a broken hiring funnel, and let the skill activate.
  5. 5Run a script on your data. Optionally run one of the three Python scripts on a JSON file of your metrics, or on the embedded sample data.

What is an agent skill and what does vpe-advisor do?

An agent skill is a folder of instructions and reference files that an AI agent loads when a request matches the skill's purpose. The skill is defined by a SKILL.md file, and it can include supporting reference documents and scripts.

vpe-advisor covers four decisions: delivery throughput (the DORA 4 metrics plus identifying where work waits), the engineering hiring funnel (conversion by stage, time-to-fill and pipeline gap), engineering team structure (squad, tribe and chapter design plus thresholds for adding a tech-lead manager), and production discipline (on-call rotation, deployment cadence and postmortem culture). The production discipline area is reference material only.

  • Delivery throughput: deployment frequency, lead time for changes, mean time to recovery and change failure rate, with bottleneck identification.
  • Hiring funnel: conversion from sourcing through screen, onsite and offer, plus time-to-fill and pipeline gap.
  • Team structure: squad, chapter and tribe design, and tech-lead manager trigger thresholds.
  • Production discipline: on-call, deployment cadence and postmortem culture, as reference guidance.
  • Not included: architecture and technology choices, which belong to a CTO-style role, and live incident response.

How do you install the vpe-advisor skill?

Download the ZIP for c-level-advisor/skills/vpe-advisor from the skill's page on the directory. Unzip it into your personal skills folder at ~/.claude/skills/, or into a project's .claude/skills/ folder if you want the skill only for that project.

If you use the Claude apps instead, open Settings, go to Capabilities, then Skills, and upload the skill there. Either way the folder should contain SKILL.md, the references folder and the scripts folder so the agent can find everything it needs.

  • Personal install: ~/.claude/skills/ (available across your projects).
  • Project install: .claude/skills/ inside the project folder (available only there).
  • Claude apps: Settings → Capabilities → Skills, then upload.
  • The skill is MIT licensed.

What does the skill need to run?

The reasoning parts of the skill need nothing beyond the agent. The three Python scripts need Python 3. They use only the standard library, so there are no packages to install.

The scripts can run on your own metrics as JSON files, or on the embedded sample data if you have no files yet. The skill runs scripts but does not use the network.

  • Python 3 (standard library only).
  • Your own metrics as JSON files, or the built-in sample data.
  • Runs scripts: yes. Uses network: no.

How do you trigger the skill with natural requests?

You do not need to type a command. Describe the problem in plain language. The skill is meant to activate when sprint velocity is dropping, engineering hiring is broken, team structure is unclear, or you are deciding when to add a tech-lead manager.

To run a calculation on your own numbers, you can point the scripts at a JSON file. The three commands below come from the skill's quick start. Replace path/to/... with your file.

  • 'Our sprint velocity has dropped for three sprints. Where is the work waiting?'
  • 'Our last four hires took 90 days and we keep losing candidates after onsite. Check our hiring funnel.'
  • 'We have 25 engineers and no clear structure. Should we add a tech-lead manager yet?'
  • 'Review our on-call rotation and postmortem process.'
python scripts/delivery_throughput_analyzer.py python scripts/delivery_throughput_analyzer.py path/to/sprint_metrics.json python scripts/eng_hiring_funnel_calculator.py python scripts/eng_hiring_funnel_calculator.py path/to/funnel.json python scripts/eng_team_structure_designer.py python scripts/eng_team_structure_designer.py path/to/team.json

What kind of answers should you expect?

For delivery, the skill compares your numbers with DORA performance bands and identifies the longest segment of cycle time, such as the PR review queue or deploy gates. For hiring, it computes conversion at each stage and points to the leakiest one, since the skill's stance is that 'we can't find good engineers' usually means the funnel has a specific problem.

For team structure, the skill recommends a shape based on headcount and work-stream complexity, and it names thresholds for adding a tech-lead manager. The guidance treats team structure as tied to architecture, so it suggests coordinating with a CTO-style role on boundaries.

  • Delivery: DORA verdict plus the top bottleneck and a suggested fix.
  • Hiring: conversion per stage, time-to-fill and pipeline gap.
  • Team structure: a recommended structure and manager-trigger status.
  • Use the DORA metrics as team-level measures, not individual performance scores.

What are the safety limits and tips?

The scripts run locally and the skill does not use the network, but they do execute code, so review any scripts before you run them in a shared or sensitive environment. The skill is reference material and not an on-call tool, so do not use it during a live outage.

Do not use the skill for individual performance evaluation or firing decisions. The source warns that a generic funnel or DORA heuristic could be unfair to a specific person. Engineers also game whatever is measured, so keep the metrics team-level.

  • Do not use it for live incident response.
  • Do not use it for architecture or technology-choice decisions.
  • Do not use it to judge or fire individuals.
  • Start with the embedded sample data to see the output format before using your own numbers.
  • Give the agent honest, complete numbers: a funnel with missing stages will produce a misleading verdict.

What people use it for

  • A startup founder checks whether deployment frequency and lead time have slipped over the last quarter and finds the top bottleneck.
  • A recruiting lead uses the funnel calculator to see whether the drop happens at the technical interview or after onsite.
  • A founder with 25 engineers decides whether the team is ready for a tech-lead manager, using the manager-trigger thresholds.
  • An engineering manager reviews whether the on-call rotation is spread across the team or concentrated on three people.

Frequently asked questions

Is the vpe-advisor skill free?

The skill is published under the MIT license, as stated on its page. The license is listed as MIT in the skill's metadata, so check the license file in the ZIP for the full terms.

What do I need to run vpe-advisor?

You need an agent that supports agent skills and Python 3 if you want to run the scripts. The scripts use only the standard library, and they can run on the built-in sample data without any files of your own.

Where do I put the vpe-advisor skill folder?

Put it in ~/.claude/skills/ for personal use or in .claude/skills/ inside a project for project-only use. In the Claude apps, upload it under Settings, then Capabilities, then Skills.

Does vpe-advisor replace a CTO?

No. The skill is explicitly not a CTO skill. A CTO-style role owns architecture and what to build, while vpe-advisor covers delivery operations, hiring, team structure and how the team ships.

Can vpe-advisor help during a production outage?

No. The skill is reference material, not an on-call tool, and it is not meant for real incident response during a live outage.

How does it compare with other engineering-metrics tools?

The source does not compare vpe-advisor with other tools, so this guide cannot make that claim. Its distinct features are the three Python scripts for arithmetic on your numbers and the explicit split from architecture decisions.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use vpe-advisor if you are a startup founder or engineering manager whose sprint velocity or release frequency is slipping, or if you want a structured read on DORA metrics or your hiring funnel. Pick something else for architecture decisions, live incident response, or any judgement about an individual's performance.

Download the skill