layers-observed-behaviour
Skill307 starsChanged 6 months agoMITView on GitHub
Techniques for planning user research and synthesising it into grounded, confidence-rated findings about what users actually do
3 other projects use a version of this, 1 of them edited.
Adopted by
Most recently taken up by
- davidanthonyn/claude-skillsedited, first seen Oct 20, 2026
- pengGgxp/skillsword for word, first seen Oct 11, 2026
- miovrag/claude-skillsword for word, first seen Jun 12, 2026
Other versions3 projects use a version of this.
Word for word · 2
- miovrag/claude-skills★ 1100% same
- pengGgxp/skills★ 1100% same
Edited · 1
- davidanthonyn/claude-skills★ 016% same
Wrote their own version · 0
None found.
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 repository8 other files it gives its agents.
- layers-conceptual-modelSkill
- layers-domainSkill
- layers-interaction-flowSkill
- layers-introSkill
- layers-orientSkill
- layers-product-strategySkill
- layers-surfaceSkill
- layers-user-needsSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 32
Ground in something seen or heard,: not in team beliefs.
≈ “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.
Files
/layers-observed-behaviour
Assumes /layers-intro has been loaded. This skill is a library of techniques, not a script — see "How to use these skills" there.
The observed behaviour layer is the closest we can get to reality — what users actually do, not what we think they do or wish they would. Everything above it is interpretation; this layer is the source.
It splits into two situations. Detect which applies and say so:
- Plan — no research yet; design a study.
- Synthesise — research material exists; make sense of it.
With partial research, synthesise what exists first, then plan to fill the gaps.
The decisions this layer makes
- What specific questions we most need to answer about our users
- What evidence already exists, and how reliable it is
- How to gather what's missing
- What patterns hold with confidence vs. what remains assumption
Disciplines — what keeps observation honest
- Stay close to raw data. Observations should be specific and near the source — what users said, did, felt — not summarised into conclusions.
- Ground in something seen or heard, not in team beliefs.
- Mark confidence: observed / inferred / assumed. If you mark something observed, the verbatim that supports it should be quotable in the same note — an observed claim with no quotable evidence is really inferred.
- Name research gaps explicitly rather than papering over them.
- Workarounds are signal. A need real enough to motivate improvisation is a strong one.
Techniques
To plan a study
| Technique | Use it when |
| Define the learning goal | Always start here. Push past "understand users better" to 2–3 specific questions — "what triggers someone to refer a friend, and what makes them hesitate." |
| JTBD interviews | Understanding triggers, motivations, anxieties. Interview about a real past experience, not hypotheticals. Guide: opening ("tell me about the last time you…"), timeline (what triggered it, what you tried), motivations (what you hoped, what worried you), closing. |
| Contextual inquiry / observation | What users say differs from what they do — watch real work for tacit behaviour. |
| Diary studies | Behaviour is distributed over time or infrequent — users self-report as events occur. |
| Support ticket / review analysis | Existing product with accumulated signal — pain points at scale without recruiting. |
| Analytics review | What users do (not why). Complements qualitative; doesn't replace it. |
| Usability observation | Where people struggle or succeed with an existing product. |
For interviews, plan synthesis up front: one observation per note, tagged with the question it speaks to, raw quotes over summaries. (6–10 qualitative interviews usually reach saturation.)
To synthesise material
| Technique | Use it to |
| Extract observations | Pull out concrete things users said, did, or felt — no interpretation yet. From memory, prompt: most surprising thing? what recurred? what did they struggle with unexpectedly? |
| Pattern grouping | Group observations by recurring situations, common motivations, shared anxieties, and workarounds. |
| Candidate job stories | When [situation], I want to [motivation], so I can [outcome]. Check the "When" is specific and the "want" is a motivation not a solution; mark confidence. |
| Gap-flagging | What do the observations not yet answer? These become a follow-up Plan session. |
Working with the designer
First find out what exists — interviews, recordings, tickets, analytics — and state the mode. Listen for nouns (candidate domain objects) and the natural language users use; that feeds the domain layer.
Offer the technique that fits: in Plan, the method matched to the learning goal; in Synthesise, extraction → patterns → candidate stories. Do the next useful thing, not a full battery.
Capture only the residue — key raw observations, the patterns with their supporting evidence, candidate job stories with confidence ratings, and the named research gaps.
Candidate job stories are ready to refine at /layers-user-needs.
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.