kpi-definition
Skill1 starsChanged 5 months agoMITView on GitHub
Define the success metrics (KPIs) and guardrails for an initiative.
In this repository7 other files it gives its agents.
- backlog-scrubSkill
- golden-testsSkill
- prd-writingSkill
- press-releaseSkill
- user-storiesSkill
- weekly-updateSkill
- yearly-kpiSkill
The file
Files
KPI definition
Turn a fuzzy goal into measurable success. Be ruthless about picking few, good metrics over many vanity ones.
Method:
- Name the outcome — the real behavior change or business result you want (not an output like "ship feature X").
- Pick the North Star for this work — the single metric that best captures the value delivered to users. Define it precisely (numerator, denominator, time window, population).
- Add input metrics — 1–3 leading indicators you can move week to week that should drive the North Star. These are what the team actually steers on.
- Set guardrails — metrics that must NOT get worse (latency, churn, support tickets, cost). A win that breaks a guardrail isn't a win.
- Baseline & target — current value, the target, and the timeframe. State the assumption behind the target.
- Instrumentation — what events/data are needed to measure this, and whether
they exist yet. Flag gaps as
prove-stage work.
Output a small table:
| Metric | Type (NS / input / guardrail) | Definition | Baseline | Target | Source |
Sanity checks:
- Could someone game this metric without helping users? If yes, fix it or pair it with a guardrail.
- Is it measurable now, or is building the measurement itself a task to log?
Record the chosen KPIs as a note on the ledger item (and in the PRD's metrics section).
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.

