Skip to content
Get packs

Guide · Agent skill

How to use the CPO Review agent skill (cpo-review)

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

The cpo-review skill is a set of instructions that tells an AI agent to review a product plan the way a chief product officer would. It asks six forcing questions covering the job a feature is hired to do, the North Star link, retention evidence for product-market fit, RICE priority, opportunity cost, and kill criteria. After this guide you will know what the skill does, how to install it, how to trigger it with a natural request, and what it needs before the output is worth trusting.

CPO Review Forcing Questionsalirezarezvani/claude-skills · MIT

Quick start

  1. 1Download the ZIP. Download the cpo-review ZIP from the directory page and unzip it.
  2. 2Choose an install location. Place the skill folder in ~/.claude/skills/ for personal use, or in .claude/skills/ inside a project. In the Claude apps, upload it under Settings → Capabilities → Skills.
  3. 3Install Python and the referenced scripts. Make sure Python is available and that the c-level-advisor and product-manager-toolkit scripts are installed under the plugin path the skill references.
  4. 4Gather your inputs. Collect the plan, the target user, and any retention or usage data you have. Without retention data, the PMF question cannot be answered properly.
  5. 5Trigger the review. Run /c-level-agents:cpo-review with your plan, or describe the decision in plain language.
  6. 6Read the verdict. Check each section of the report, paying close attention to the PMF Signal and Kill Criteria before acting on the SHIP, SHARPEN, or KILL verdict.

What the cpo-review skill makes the agent do

An agent skill is a folder containing a SKILL.md file. The file tells the agent when to use a method and how to structure its answer. This skill is for challenging a product plan before resources are committed. Its description says it is for deciding whether to kill a feature, committing a quarter's roadmap, or claiming product-market fit (PMF).

When you ask the agent to review a plan, it works through six questions: what job the feature is hired to do in the user's words, how it ties to the company's North Star metric, what the retention curve shows, its RICE score, what gets cut to make room, and what result in 90 days would mean it was a mistake. It then produces a structured report ending in a SHIP, SHARPEN, or KILL verdict.

  • JTBD: state the job in the user's voice, not as a feature name
  • North Star: trace the feature to a leading, behavior-based metric
  • PMF signal: classify the retention curve as flat, smiling, or decaying
  • RICE: score Reach, Impact, Confidence and Effort
  • Opportunity cost: name the specific initiative that gets cut
  • Kill criteria: write the metric and threshold before launch

Install the cpo-review skill

Download the ZIP from the directory page and unzip it. For a personal setup, place the unzipped skill folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder instead.

If you use the Claude apps rather than a local setup, open Settings, go to Capabilities, and upload the skill under Skills.

  • Personal install: ~/.claude/skills/
  • Project install: .claude/skills/ inside your project
  • Claude apps: Settings → Capabilities → Skills, then upload

Set up the scripts it depends on

The skill's SKILL.md refers to helper scripts for scoring. The file itself only contains the prompts and report template, so the scripts must be present separately. The skill references the c-level-advisor and product-manager-toolkit scripts from the same repository, installed under the plugin path. The scoring script that the RICE step calls is in the product-manager-toolkit folder, and the PMF and portfolio scripts are under the c-level-advisor vendor path.

You need Python installed, since the referenced scripts are Python files. The skill does not supply retention data, so you must bring your own for a real PMF assessment.

  • Python is required
  • c-level-advisor and product-manager-toolkit scripts must be installed under the plugin path
  • Retention or usage data is your responsibility
python product-team/skills/product-manager-toolkit/scripts/rice_prioritizer.py

Trigger the skill with a natural request

The skill's documented command is /c-level-agents:cpo-review followed by the plan. You can also ask in plain language. The description lists the situations it fits: committing a quarter's roadmap, deciding whether to kill a feature, or claiming PMF. Give the agent the plan, the target user, and any numbers you have. The more concrete the input, the more useful the six answers.

Good triggers describe a decision. Examples: 'Review our Q3 roadmap with the cpo-review skill before we commit,' or 'Should we kill the bulk export feature? Run a CPO review using our usage data.'

  • Before a quarterly roadmap commitment
  • Before adding more than three features to one release
  • When retention is flat or declining
  • When the team is debating whether to build something new
/c-level-agents:cpo-review <plan>

Read the output report

The skill returns a markdown report with fixed sections: JTBD, North Star Link, PMF Signal, Score, Cut List, Kill Criteria (90 days), and Verdict. The verdict is one of three labels: SHIP, SHARPEN, or KILL, shown with green, yellow, or red markers.

Treat the PMF Signal section with care. The skill's rule is that a flat or smiling retention curve is a PMF signal and a decaying one is not, and that survey enthusiasm is not a signal. If your data does not contain a retention curve, the PMF section is where the agent is most likely to guess, so check it first.

  • JTBD: one sentence in user voice
  • North Star Link: metric moved and expected change
  • PMF Signal: curve shape and cohort sample size
  • Score: RICE number and rank in queue
  • Cut List: what is cut and why
  • Kill Criteria: metric, threshold, and action if missed
  • Verdict: SHIP, SHARPEN, or KILL

Safety notes and what the skill can and cannot do

This skill runs scripts. It does not use the network, according to its listing. Read the scripts before running them, especially if you install them from a plugin path, and run them on data you are comfortable processing locally.

The skill does not produce market research or executive judgement. Its output is only as good as the inputs you give it.

  • Runs scripts: yes
  • Uses network: no
  • Licence: MIT
  • Retention data is not supplied by the skill

What people use it for

  • Pressure-testing a quarterly roadmap before the team commits engineering time
  • Deciding whether to kill an underused feature and naming what it frees up
  • Challenging a founder's claim that a product has PMF by checking for a retention curve
  • Writing 90-day kill criteria for a new feature before launch
  • Checking whether a proposed release with more than three features is too large

Frequently asked questions

How do I install the cpo-review agent skill?

Download the ZIP from the directory page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. In the Claude apps, you can instead upload it under Settings → Capabilities → Skills.

What do I need to run the cpo-review skill?

You need Python, the c-level-advisor and product-manager-toolkit scripts from the same repository installed under the plugin path, and retention or usage data for a real PMF assessment. The skill does not supply the retention data.

Is the cpo-review skill free?

The skill is listed under the MIT licence. The source does not state any price, so check the repository for licence terms.

Does cpo-review need an internet connection?

The listing says it does not use the network. It does run scripts, so review them before running.

Can the cpo-review skill tell me if my product has product-market fit?

Not on its own. The skill checks the retention curve shape, but it cannot produce one. If you lack retention or usage data, the agent may guess or overstate confidence, so do not rely on the PMF result without real data.

What is the difference between SHIP, SHARPEN, and KILL?

These are the three verdict labels the skill returns at the end of its report. SHIP means the plan goes ahead, SHARPEN means it needs revision, and KILL means it should be stopped, based on the six-question analysis.

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 cpo-review if you are a product manager, founder, or leader who wants a structured challenge before committing a roadmap or killing a feature, and you have real retention or usage data. Skip it if you have no retention data or need real market research, because the PMF question and the overall verdict will only be as sound as the inputs you provide.

Download the skill