Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add varunk130/ai-customer-discovery-skills --skill north-star-metric-findergit clone --depth 1 https://github.com/varunk130/ai-customer-discovery-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/varunk130/ai-customer-discovery-skills/north-star-metric-finder)<a href="https://agentmods.dev/skills/varunk130/ai-customer-discovery-skills/north-star-metric-finder"><img src="https://agentmods.dev/badge/skills/varunk130/ai-customer-discovery-skills/north-star-metric-finder/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/varunk130/ai-customer-discovery-skills/north-star-metric-finder"><img src="https://agentmods.dev/badge/skills/varunk130/ai-customer-discovery-skills/north-star-metric-finder.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00086 | $0.00969 |
| Opus 5 | $0.00043 | $0.00485 |
| Sonnet 5 | $0.00017 | $0.00194 |
| Haiku 4.5 | $0.00009 | $0.00097 |
Grade A, and why
north-star-metric-finder scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.
North Star Metric Finder
Identify a candidate North Star Metric (NSM) that satisfies five strict criteria, then map the 3-5 input metrics that drive it. The output is a defensible recommendation, not a brainstormed list.
Core Principle
A North Star Metric is a forcing function, not a slogan. Adopting the wrong one quietly skews two years of roadmap decisions. The five criteria below exist to eliminate candidates that look reasonable but fail under stress.
The Five Criteria
| Criterion | The Question | Failure Example |
|---|---|---|
| Value-capturing | Does this metric only go up when customers receive real value? | "Logins per week" - easy to game with notification spam |
| Predictive | Does this metric lead revenue and retention by ≥1 quarter? | "MRR" - lagging, not leading |
| Actionable | Can the team move this metric with deliberate actions? | "NPS" - moves slowly, hard to attribute changes |
| Understandable | Can every employee state the metric and why it matters in one sentence? | "Weekly active accounts with ≥3 successful API calls in their primary workspace" |
| Singular | Is this a single metric, not an index of three? | A weighted composite hides which input is broken |
A candidate must pass all five. Three out of five is not a NSM - it's a useful KPI.
Output
Save to outputs/nsm-[product]-[YYYY-MM-DD].md
- Recommended NSM + the five-criteria pass/fail evaluation
- Runner-up candidates with the criterion they failed and why
- Input metrics (3-5) that drive the NSM, with the direction and expected magnitude
- Counter-metrics that prevent gaming (e.g., NSM ↑ + churn ↑ = bad)
- Cadence - how often to review the NSM and re-validate it
Process
Step 1: Generate Candidates
I'll ask:
"What does your product do for the customer? What's the moment they get value? Share your current top-line metrics and a one-sentence pitch."
I generate 6-10 candidate NSMs grounded in the value moment, not the business model.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 72 lines · 86 tokens per session scan A 3de21084c301
north-star-metric-finder is a skill published in the GitHub repository varunk130/ai-customer-discovery-skills (2 stars, last pushed 1mo ago), licensed MIT. It adds 86 tokens to every session and 969 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
profile-builder
Produces discovery-context.md — the single source of truth that every other skill in this pack reads. Without it, other skills will warn and offer to bootstrap inline (a 2–3 question mini-interview held in conversation, not written to disk). Running this skill first gives the best, most consistent results.
solution-architecture
Produces a discovery-grade architecture: enough detail to estimate effort, surface technical risks, and explain choices to the client — but not detailed design. Detailed design happens in delivery, not here. Where the BA isn't a tech lead, this skill is best run pair-mode with an architect; otherwise it produces vague…
insight-synthesis
Turns whatever evidence the BA managed to gather into structured insights — weighted by source quality so the team doesn't treat one SME's hunch as equal to five user interviews.
user-interviews
Direct user interviews. Run only when useraccess allows it.
discovery-handoff
When discovery ends and delivery begins — possibly with a different team. This document is the bridge. Reading it cold should be enough to start delivery work.
estimation
Agency-style estimation: effort by discipline × phase, with low / expected / high ranges, and a list of assumptions that, if violated, invalidate the estimate. The output is the heart of any commercial proposal.