agentleFS
Sign inStart free
rpraharajrpraharaj/ai-sdlc/skills/design-experiment/SKILL.md

design-experiment

Skill0 starsChanged 17 days agoMITView on GitHub

Design an experiment with a hypothesis, an existing product metric, and a stop rule. Use when the user asks for an A/B test, experiment design, hypothesis, or /design-experiment. Do not invent metrics or baselines. Do not write product code.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Write a falsifiable bet using a metric this product already has.

    line 17
  2. Distinctive:

    Do not invent dashboards, baselines, or sample sizes.

    line 19
  3. Very specific:

    If no existing metric exists, stop.

    line 34
  4. Distinctive:

    Stop rule (ship / iterate / drop)

    line 38
  5. Distinctive:

    Read AGENTS.md if present.

    line 44
  6. Distinctive:

    No secrets. No live personal data in the doc.

    line 50

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Design experiment

Write a falsifiable bet using a metric this product already has. This is not a PRD, not spec-analytics, and not implementation.

Do not invent dashboards, baselines, or sample sizes.

Is this the right skill?

SignalDo this instead
What events to emitspec-analytics
What to build if the bet is already decidedwrite-prd / shape-work
Quality targetswrite-nfr

Required inputs

  • A hypothesis the user can state (or an opportunity/PRD to extract one from)
  • A metric already in the product, PRD analytics section, or an existing dashboard the user names

If no existing metric exists, stop. Name spec-analytics only for adding events; do not invent a success number.

Human decisions (do not invent)

  • Stop rule (ship / iterate / drop)
  • Who may be in the experiment
  • Any duration not already constrained

Procedure

  1. Read AGENTS.md if present. Read the PRD or opportunity if named.
  2. Confirm the metric exists in a file or a named dashboard. Otherwise unknown and stop for the human.
  3. Find the home: docs/discover/experiment-YYYY-MM-DD-<slug>.md.
  4. Open references/experiment-template.md. Copy every heading. Write n/a when unused.
  5. Hypothesis: if we X, then metric Y (already exists) moves in direction Z, because …
  6. Guardrails: what must not get worse (existing metrics only, or unknown).
  7. No secrets. No live personal data in the doc.

Output

  • Path: docs/discover/experiment-YYYY-MM-DD-<slug>.md
  • Template: references/experiment-template.md

Quality bar

  • Metric names a source (path, dashboard this program already uses, or user-named)
  • No invented baseline
  • Stop rule is present or unknown — need human
  • Unused headings are n/a

Done

  1. Path
  2. Metric + stop rule or unknown
  3. Next skill: spec-analytics if events are missing; shape-work if they will implement a flag

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.

Check your setup, free