synthesize

Name: synthesize
Author: tuan3w

32stars2forks

How to install synthesize?

tuan3w/obsidian-vault-agent/synthesize

$npx -y skills add tuan3w/obsidian-vault-agent --skill synthesize

Installs into the current project.

Summary

Deep cross-domain synthesis engine for the vault. Triggers on "synthesize X", "what do I know about X", "connect ideas about X", "cross-domain analysis of X", "map everything in the vault about X". Searches comprehensively, surfaces non-obvious patterns and tensions, then challen

Files · 1

View on GitHub

SKILL.md

1	<Purpose>
2	Exhaust the vault's knowledge on a topic — across ALL domains — then act as
3	Luhmann's communication partner: surface contradictions the user hasn't noticed,
4	identify the hidden principle generalizing across examples, and demand original
5	synthesis. This is not a retrieval tool. It is a thinking provocation.
6
7	Science grounding: far transfer (Tempel & Frings 2024), generation effect
8	(McCurdy et al. 2020), elaborative interrogation, desirable difficulty
9	(Bjork & Bjork 2020).
10	</Purpose>
11
12	<Use_When>
13	- User wants to understand what the vault collectively says about a concept
14	- User is writing a Thought note and wants to pressure-test an idea
15	- User is preparing a talk, essay, or decision and needs cross-domain ammunition
16	- User says "synthesize X", "what do I know about X", "connect ideas about X"
17	- User wants to find contradictions or gaps in their own thinking
18	</Use_When>
19
20	<Do_Not_Use_When>
21	- User wants to process a single source note — use /process instead
22	- User wants spaced repetition review — use /recall instead
23	- User has a specific paper to analyze — use /paper instead
24	- Topic is too narrow to generate cross-domain connections (single paper, single event)
25	</Do_Not_Use_When>
26
27	<Steps>
28
29	## Stage 1: PARSE THE TOPIC
30
31	Extract the core concept(s) from the user's input. Identify:
32	- Primary term: the exact word(s) to search (e.g., "attention", "loss aversion")
33	- Synonyms / adjacent: related terms likely to appear in notes (e.g., "focus", "salience", "bias")
34	- Domain anchors: which domains likely hold relevant notes (ML, psychology, design, startup...)
35
36	If the topic is ambiguous, briefly clarify before searching (one question max).
37
38	## Stage 2: PARALLEL SEARCH — run ALL of these simultaneously
39
40	MCP search (preferred):
41	```
42	search_notes(query="PRIMARY TERM", limit=30)
43	search_notes(query="SYNONYM 1", limit=15)
44	search_notes(query="SYNONYM 2", limit=15)
45	```
46
47	Grep fallback (if MCP unavailable):
48	```
49	Grep(pattern="PRIMARY TERM", path="notes/", glob="*.md", output_mode="files_with_matches")
50	Grep(pattern="#TAG-NAME", path="notes/", glob="*.md", output_mode="files_with_matches")
51	Grep(pattern="\[\[.PRIMARY TERM.\]\]", path="notes/", glob="*.md", output_mode="files_with_matches")
52	```
53
54	Also search:
55	- Thought notes referencing this topic: `Grep(pattern="PRIMARY TERM", path="notes/thoughts/", glob="*.md")`
56	- Cross-domain appearances: search each domain folder separately to surface unexpected hits
57	- Existing Term notes: `Grep(pattern="type: term", path="notes/", glob="*.md")` then filter
58
59	Collect all matching file paths. Deduplicate. Aim for 15–25 notes across diverse domains.
60
61	## Stage 3: READ THE NOTES
62
63	Read up to 20 most relevant notes using `read_multiple_notes` or individual `read_note` calls.
64
65	Prioritize:
66	1. Thought notes (user's own synthesis — highest signal)
67	2. Term/Note notes (distilled concepts)
68	3. Paper/book/post notes with substantial content
69	4. Notes from UNEXPECTED domains (highest cross-domain value)
70
71	While reading, track:
72	- Key claims and insights
73	- Domain of each note
74	- Wikilinks that connect to the topic
75	- Any explicit contradictions or tensions
76	- Claims that could conflict with each other
77
78	## Stage 4: PRESENT THE SYNTHESIS
79
80	Output this structure (adapt headings as needed, cut any section with nothing real to say):
81
82	---
83
84	## What the vault says about [TOPIC]
85
86	### Key insights
87	[3–6 bullets. Each = one non-obvious claim distilled from multiple notes. Lead with the sharpest insight. No filler.]
88
89	### Cross-domain appearances
90	[Table or bullet list: Domain → how the concept appears there. Include only non-obvious mappings — skip domains where the concept is trivially expected.]
91
92	\| Domain \| How [topic] appears \|
93	\|--------\|---------------------\|
94	\| [domain] \| [specific claim from that note] \|
95
96	### Agreements
97	[What multiple notes converge on — the vault's "consensus". 2–4 bullets max.]
98
99	### Tensions & contradictions
100	[Where notes disagree — including tensions the user may not have noticed. THIS IS THE MOST VALUABLE SECTION. If two notes from different domains make opposing claims, surface it explicitly. Quote the note titles.]
101
102	### Gaps
103	[What's missing: questions the vault raises but doesn't answer, adjacent concepts not yet captured, domains that probably have something to say but no no

People who liked this also viewed

6 picks

Type

synthesize

32stars2forks

View on GitHub

How to install synthesize?

tuan3w/obsidian-vault-agent/synthesize

$npx -y skills add tuan3w/obsidian-vault-agent --skill synthesize

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	<Purpose>
2	Exhaust the vault's knowledge on a topic — across ALL domains — then act as
3	Luhmann's communication partner: surface contradictions the user hasn't noticed,
4	identify the hidden principle generalizing across examples, and demand original
5	synthesis. This is not a retrieval tool. It is a thinking provocation.
6
7	Science grounding: far transfer (Tempel & Frings 2024), generation effect
8	(McCurdy et al. 2020), elaborative interrogation, desirable difficulty
9	(Bjork & Bjork 2020).
10	</Purpose>
11
12	<Use_When>
13	- User wants to understand what the vault collectively says about a concept
14	- User is writing a Thought note and wants to pressure-test an idea
15	- User is preparing a talk, essay, or decision and needs cross-domain ammunition
16	- User says "synthesize X", "what do I know about X", "connect ideas about X"
17	- User wants to find contradictions or gaps in their own thinking
18	</Use_When>
19
20	<Do_Not_Use_When>
21	- User wants to process a single source note — use /process instead
22	- User wants spaced repetition review — use /recall instead
23	- User has a specific paper to analyze — use /paper instead
24	- Topic is too narrow to generate cross-domain connections (single paper, single event)
25	</Do_Not_Use_When>
26
27	<Steps>
28
29	## Stage 1: PARSE THE TOPIC
30
31	Extract the core concept(s) from the user's input. Identify:
32	- Primary term: the exact word(s) to search (e.g., "attention", "loss aversion")
33	- Synonyms / adjacent: related terms likely to appear in notes (e.g., "focus", "salience", "bias")
34	- Domain anchors: which domains likely hold relevant notes (ML, psychology, design, startup...)
35
36	If the topic is ambiguous, briefly clarify before searching (one question max).
37
38	## Stage 2: PARALLEL SEARCH — run ALL of these simultaneously
39
40	MCP search (preferred):
41	```
42	search_notes(query="PRIMARY TERM", limit=30)
43	search_notes(query="SYNONYM 1", limit=15)
44	search_notes(query="SYNONYM 2", limit=15)
45	```
46
47	Grep fallback (if MCP unavailable):
48	```
49	Grep(pattern="PRIMARY TERM", path="notes/", glob="*.md", output_mode="files_with_matches")
50	Grep(pattern="#TAG-NAME", path="notes/", glob="*.md", output_mode="files_with_matches")
51	Grep(pattern="\[\[.PRIMARY TERM.\]\]", path="notes/", glob="*.md", output_mode="files_with_matches")
52	```
53
54	Also search:
55	- Thought notes referencing this topic: `Grep(pattern="PRIMARY TERM", path="notes/thoughts/", glob="*.md")`
56	- Cross-domain appearances: search each domain folder separately to surface unexpected hits
57	- Existing Term notes: `Grep(pattern="type: term", path="notes/", glob="*.md")` then filter
58
59	Collect all matching file paths. Deduplicate. Aim for 15–25 notes across diverse domains.
60
61	## Stage 3: READ THE NOTES
62
63	Read up to 20 most relevant notes using `read_multiple_notes` or individual `read_note` calls.
64
65	Prioritize:
66	1. Thought notes (user's own synthesis — highest signal)
67	2. Term/Note notes (distilled concepts)
68	3. Paper/book/post notes with substantial content
69	4. Notes from UNEXPECTED domains (highest cross-domain value)
70
71	While reading, track:
72	- Key claims and insights
73	- Domain of each note
74	- Wikilinks that connect to the topic
75	- Any explicit contradictions or tensions
76	- Claims that could conflict with each other
77
78	## Stage 4: PRESENT THE SYNTHESIS
79
80	Output this structure (adapt headings as needed, cut any section with nothing real to say):
81
82	---
83
84	## What the vault says about [TOPIC]
85
86	### Key insights
87	[3–6 bullets. Each = one non-obvious claim distilled from multiple notes. Lead with the sharpest insight. No filler.]
88
89	### Cross-domain appearances
90	[Table or bullet list: Domain → how the concept appears there. Include only non-obvious mappings — skip domains where the concept is trivially expected.]
91
92	\| Domain \| How [topic] appears \|
93	\|--------\|---------------------\|
94	\| [domain] \| [specific claim from that note] \|
95
96	### Agreements
97	[What multiple notes converge on — the vault's "consensus". 2–4 bullets max.]
98
99	### Tensions & contradictions
100	[Where notes disagree — including tensions the user may not have noticed. THIS IS THE MOST VALUABLE SECTION. If two notes from different domains make opposing claims, surface it explicitly. Quote the note titles.]
101
102	### Gaps
103	[What's missing: questions the vault raises but doesn't answer, adjacent concepts not yet captured, domains that probably have something to say but no no

People who liked this also viewed

6 picks

Type