demo-data
Skill2 starsChanged 49 days agoMITView on GitHub
List, keep, separate, or delete the demo dataset so it's never mistaken for real research. Use on /demo-data or when asked to clean up/remove/separate demo content.
In this repository31 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- affinitySkill
- ai-personaSkill
- analytics-syncSkill
- backlogSkill
- catalogSkill
- cold-startSkill
- contradictionsSkill
- dovetail-syncSkill
- exportSkill
- extract-findingsSkill
- feature-panelSkill
- forget-participantSkill
- graph-lintSkill
- interview-guideSkill
- journey-mapSkill
- opportunity-treeSkill
- persona-avatarSkill
- persona-querySkill
- persona-talkSkill
- persona-voiceSkill
- persona-workshopSkill
- prdSkill
- prioritizeSkill
- researcherSkill
- retentionSkill
- source-syncSkill
- subject-accessSkill
- welcomeSkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 2 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 28
Write the marker, done. Note for the user: the demo remains the best live reference of a fully-wired graph — and persona-talk's ⚠️ preamble …
- Very specific: line 45
Add a one-line Demo/README.md: *"Example dataset (Spotify) — illustrative, demo: true everywhere, not real research.
- Very specific: line 46
Run python3 scripts/graph_lint.py (it scans only top-level folders, so separated demo content drops out of lint noise — by design) and fix …
- Distinctive: line 48
Write the marker; offer a local commit ("Separate demo dataset into Demo/").
- Very specific: line 54
Update DEMO.md (points at content that no longer exists → replace body with a note that the example set was removed and the walkthrough …
- Distinctive: line 55
Run graph-lint; write the marker.
- Very specific: line 59
Never mix decisions: one run = one of keep/separate/delete, applied consistently to the whole demo set — a half-moved demo is worse than …
≈ “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
demo-data — Keep, separate, or delete the example dataset
Every example file carries demo: true in frontmatter (badge in the app, ⚠️ preamble in persona-talk). That's machine-clear, but as real research accumulates, humans skimming folders can still mix demo and real files. This skill executes a one-time decision (revisitable any time) and records it in .claude/demo-decision.local so nothing nags again.
App shortcut: the browser app's Settings page has a "Delete demo files" button — it mechanically removes every demo: true entity file from disk (never touching unflagged files in the same folders) and suppresses the app's embedded example preview. The delete mode below remains the full cleanup: demo avatars, DEMO.md, demo rows in Research backlog.md/Participants.md, README's example-set paragraph.
Trigger
/demo-data ← show status + the three options
/demo-data status ← just list what's demo vs real, per folder
/demo-data keep ← keep in place (badges already separate it) — records the decision
/demo-data separate ← move all demo content under Demo/ (mirrored folders)
/demo-data delete ← remove all demo content (destructive — double confirmation)Status (always run first)
Count demo: true files per folder vs files without the flag. Say plainly which state the repo is in: pure demo (no real data yet — recommend keep, it's the working reference), mixed (recommend separate), or real-only-plus-demo-leftovers (separate or delete).
keep
Write the marker, done. Note for the user: the demo remains the best live reference of a fully-wired graph — and persona-talk's ⚠️ preamble + the app's Demo badge keep it unmistakable.
separate — move under Demo/, mirrored structure
Target layout preserves relative depth, so the demo's internal ../Signals/… links keep working unchanged:
Demo/Personas/ Demo/Signals/ Demo/Evidence/ Demo/Archetypes/
Demo/Ideas/ Demo/Competitors/ Demo/Transcripts/- Plan first: list every
demo: truefile and its destination; get one approval for the whole move (this is a mechanical relocation, not a graph rewiring — the CLAUDE.md link-approval rule's "batch trivial fixes" case). git mveach file (history preserved).- Special cases, handled explicitly:
Product Context.md— skills depend on this root path. Archive the demo content toDemo/Product Context (Spotify demo).mdand reset the root file to a skeleton for the user's own product (offer/welcomeStep 4 to fill it).Research backlog.md/Participants.md— move demo rows to a clearly-marked "Demo examples" section at the bottom of each file (or delete the rows if the user prefers); the files themselves stay.DEMO.md— update its paths toDemo/Personas/Emma.mdso the demo walkthrough still works.
- Add a one-line
Demo/README.md: "Example dataset (Spotify) — illustrative,demo: trueeverywhere, not real research. Talk to it:/persona-talk Demo/Personas/Emma.md." - Run
python3 scripts/graph_lint.py(it scans only top-level folders, so separated demo content drops out of lint noise — by design) and fix any real-data links the move surfaced. - Persona cards in
.claude/cache/invalidate automatically (graph hash changed). - Write the marker; offer a local commit ("Separate demo dataset into Demo/").
delete — destructive, slow down
- Show the full list of what will be removed (files + demo rows).
- Commit the current state first ("Snapshot before demo-data delete") so it's recoverable from history — then, after an explicit second confirmation,
git rmthe demo files and strip demo rows. - Update
DEMO.md(points at content that no longer exists → replace body with a note that the example set was removed and the walkthrough requires the template's original data) andREADME.md's example-set paragraph. - Run graph-lint; write the marker; offer a local commit.
Rules
- Never mix decisions: one run = one of keep/separate/delete, applied consistently to the whole demo set — a half-moved demo is worse than either state.
- Real data (no
demo:flag) is never touched by this skill, whatever the mode. - The marker
.claude/demo-decision.local(gitignored) stores the choice + date;/demo-datacan always be re-run later to change the decision.
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.

