vault-emerge
Skill0 starsChanged 7 months agoMITView on GitHub
Surface unnamed patterns from recent vault activity. Reads N days of notes and finds recurring themes you haven't explicitly named. The vault notices what you don't.
In this repository9 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- llm-wikiSkill
- vault-challengeSkill
- vault-connectSkill
- vault-graduateSkill
- vault-ingestSkill
- vault-saveSkill
- vault-worldSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 5 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
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 18
Read daily notes from last N days.
- Distinctive: line 19
Read project notes modified in last N days.
- Distinctive: line 20
Read research notes modified in last N days.
- Distinctive: line 21
Read other domain folders modified in last N days.
- Distinctive: line 63
After vault-health finds many orphan notes
≈ “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
/vault-emerge [days=30]
Find patterns hiding in your vault that you haven't named yet.
Steps
- Read
_CLAUDE.mdat vault root for folder map. - Collect raw material -- spawn parallel reader agents (one per major folder):
- Read daily notes from last N days. Extract: topics, decisions, blockers, mood signals.
- Read project notes modified in last N days. Extract: status changes, recurring blockers, tools mentioned.
- Read research notes modified in last N days. Extract: themes, frameworks, open questions.
- Read other domain folders modified in last N days. Extract: patterns, failures, workarounds.
- Pattern detection -- look for:
- Recurring topics appearing 3+ times across unrelated contexts
- Unnamed workflows -- action sequences that keep happening but have no name
- Shifting priorities -- what's getting more attention over time? Less?
- Unresolved tensions -- contradictions between different notes
- Emerging expertise -- domains where depth is growing but not acknowledged
- Abandoned threads -- topics that appeared briefly then vanished
Present findings as named patterns:
## Emerged Patterns (last N days) ### 1. [Pattern Name] -- [one-line description] **Evidence:** [3+ specific notes with dates] **Frequency:** [how often, across what contexts] **So what:** [why this matters -- actionable implication] **Suggestion:** [formalize as protocol? create concept note? investigate deeper?]- Rank by signal strength -- more evidence + more diverse sources = higher rank.
- Ask: "Any of these worth naming? I can create concept notes for them."
- Append to log.md:
## [YYYY-MM-DD] emerge | [N] patterns surfaced from [days] days
Rules
- All output in the user's language (from
_CLAUDE.mdLanguage section). Translate ALL template headings and labels. E.g., Chinese user sees "## 浮现的模式", "### 1. [模式名]", "证据:", not English. Pattern names should be in the user's language. Evidence quotes stay in original language. Search in all languages present in the vault. - Minimum 3 independent data points to call something a pattern
- Cross-domain signals are gold -- same pattern in two unrelated areas is more interesting
- Name patterns concisely -- if you can't name it in 5 words, you haven't understood it
- Don't state the obvious -- "user works on multiple projects" is not a pattern
- Include counter-evidence -- if a pattern has exceptions, mention them
When to Suggest
- Weekly review time
- User feels scattered or directionless
- Before planning sessions
- After vault-health finds many orphan notes
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.

