product-metrics
Skill5 starsChanged 28 days agoMITView on GitHub
Build the measurement layer of a product — one candidate North Star, the tree of levers that hold it up with where each is measured, one metric per active hypothesis with its value today or a gap with an owner, and the antimetrics with their alarm thresholds — and write the Metric Brief. Manually triggered, or entered when the metric gate of product-strategy does not pass.
3 other projects use a version of this.
Adopted by
Most recently taken up by
- AntonioMellies/claude-code-skillstheir own version, first seen Nov 16, 2026
- okou-ai/okou-skillstheir own version, first seen Nov 11, 2026
- vm0-ai/vm0-skillstheir own version, first seen Oct 8, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- okou-ai/okou-skills★ 81same name
- vm0-ai/vm0-skills★ 81same name
- AntonioMellies/claude-code-skills★ 2same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository12 other files it gives its agents.
- cpoSkill
- ctoSkill
- grillSkill
- ideation-engineSkill
- insight-engineSkill
- market-researchSkill
- prdSkill
- product-builder-resolverSkill
- product-strategySkill
- synthetic-usersSkill
- ux-researchSkill
- visual-designerSkill
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 5 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
5
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 31
Before asking: if the loaded product node has dated research from product-strategy (a Discovery Brief with hypotheses), read it — the …
- Distinctive: line 33
If there is no prior research, they are asked directly.
- Distinctive: line 46
Every lever carries where it is measured (which event, which table, which source — the where, never how to instrument it: that belongs to …
- Distinctive: line 69
After every proposed metric (North Star, lever or hypothesis metric), this question is applied:
- Distinctive: line 72
If there is no answer, or the answer is another metric ("it changes how we see growth"), it is a vanity metric: it is said outright — "this …
- Distinctive: line 105
If the interview detected a vanity metric along the way, the Brief leaves it written anyway —with the label "(vanity: it does not change …
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
product-metrics — the Metric Brief
The gap the old pipeline acknowledged without filling: product-strategy demands that "with no
metric, the problem is not clear", but no tool helped build that metric — it handed off to a metrics
coach that never existed. This is that station: it defines the candidate North Star, builds the tree
of levers that hold it up, and gives every active hypothesis a number, a way of measuring it and a
value as of today.
When it is invoked
| When | From / to |
The metric gate of product-strategy does not pass ("with no metric, the problem is not clear") | Entry from product-strategy (.agents/skills/product-strategy/SKILL.md) |
| The Metric Brief is written | Exit toward prd (.agents/skills/prd/SKILL.md) — the strategic context/ of the node is completed with the metric already defined |
| With no prior gate | Invocable on its own, over the product node already loaded |
Both neighboring stations are already installed in this pack: naming them here is real, not invented — unlike the old reference to a metrics coach with neither spec nor owner.
The interview
It is run with the pressure method of .agents/skills/grill/SKILL.md#method — cited by path, never copied.
The five ingredients (counter-question with an example, pressure capped at 1-2 attempts, escape
hatches, evidence hierarchy, every gap recorded) are applied exactly as described there.
Before asking: if the loaded product node has dated research from product-strategy (a Discovery
Brief with hypotheses), read it — the active hypotheses come from there, they are not asked again.
If there is no prior research, they are asked directly.
Four steps, in this order:
1. Candidate North Star
A single candidate North Star metric, with the reason in one line: which product decision it moves, not why "it matters". Faced with an answer that names no decision ("because it measures success", "because it is the one everybody looks at"), grill returns it with a counter-question.
2. Metric tree
The North Star on top, its levers below — the variables that, if they move, move the North Star. Every lever carries where it is measured (which event, which table, which source — the where, never how to instrument it: that belongs to the body of the user's product, not to this tool). A lever with no where-it-is-measured is a gap, not a complete line.
3. Metric per active hypothesis
For each active hypothesis (the ones product-strategy brings, or the ones raised here if there is
no prior research): which number validates or refutes it, measured how, against what value as of
today.
Hard rule: no value with a source, no data. A value as of today that the operator cannot cite
with its origin (a dashboard, a measurement, a dated figure) is not written as if they could: it is
recorded as a gap, with who closes it — the same gap format grill uses. Inventing the number so the
row looks complete is the same incomplete scan presented as complete that the rest of the system
forbids.
4. Antimetrics
What should not get worse while the North Star improves, with the threshold that triggers the alarm — a number or a verifiable condition, never "if it gets much worse".
The vanity metric
After every proposed metric (North Star, lever or hypothesis metric), this question is applied: which decision does this number change, in whichever direction it moves?
If there is no answer, or the answer is another metric ("it changes how we see growth"), it is a vanity metric: it is said outright — "this is a vanity metric: it changes no decision" — and the concrete decision the metric should move is requested before accepting it into the Brief. It is neither discarded in silence nor accepted to avoid interrupting the interview.
The Metric Brief
It closes by writing the deliverable as dated research of the loaded product node, in:
research/<YYYY-MM-DD>-metric-brief.mdDated, never overwritten — the same criterion the research folder of this repo already uses. The
path falls inside content: */products/*/research/*.md line of the brain's own tree.md: no new
glob and no resolver row are needed. Without a product node, this same file —
<YYYY-MM-DD>-metric-brief.md — is what lands in the current folder instead.
The Brief carries these four sections, in this order, each with its literal heading:
## Candidate North Star
## Metric tree
## Metric per hypothesis
## Antimetrics- Candidate North Star: the metric and the reason in one line (which decision it moves).
- Metric tree: the North Star on top, each lever on one line with its where-it-is-measured.
- Metric per hypothesis: one entry per active hypothesis — which number, measured how, value as of today or a gap with an owner.
- Antimetrics: each one with its alarm threshold.
If the interview detected a vanity metric along the way, the Brief leaves it written anyway —with the label "(vanity: it does not change [the decision that was asked for and never got an owner, if it was left unclosed])"— and never deletes it in silence.
What this deliverable does not claim
The Metric Brief neither creates nor demands a new canonical file of the product node: it is dated
research, it adds to what is already there, and it neither replaces nor forces re-versioning a live
file of the context/. If a metrics canonical file were ever needed (a live metrics.md that gets
overwritten), that decision is approved by the operator — not by this tool.
Destination
The skill resolves the destination before producing the deliverable, never after. Three cases, two destinations:
- A product node is loaded: the deliverable goes to
research/<YYYY-MM-DD>-<name>.mdof that node. - There is a brain, but no product node — no workspace yet, or a workspace without one: the deliverable is written to a file in the current folder.
- There is no brain at all: the deliverable is written to a file in the current folder.
With no product node, the file is written to the current folder anyway. The closing message names only the path it just wrote and stops there — that is the normal way this skill ends, not an anomaly to qualify.
This skill never asks which destination to use, and it never invents a third one.
The rest of the pack installs one skill at a time. Look at .agents/skills/ first and then .claude/skills/ (a skill present in both
counts once) and offer only the ones that are not there — once per session: the first time this skill closes in the session,
never again on a later close of the same or another deliverable:
grill—npx skills add pedroromeroluna/ai-first-product-skills --skill grillprd—npx skills add pedroromeroluna/ai-first-product-skills --skill prdproduct-strategy—npx skills add pedroromeroluna/ai-first-product-skills --skill product-strategy
The whole pack at once: npx skills add pedroromeroluna/ai-first-product-skills. That command installs what the pack offers;
anything listed above it is installed by naming it.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

