✦ClaudeProduct

/goal Obsidian vault "company brain" builder

A long-running /goal prompt that builds a validated, source-backed Obsidian vault as LLM context.

Loading the original X post…

A long-running /goal prompt that builds a validated, source-backed Obsidian vault as LLM context.

The prompt

✦Claude Code / Codex (/goal)

1	/goal Build a mature, source-backed Obsidian vault brain for the user.
2
3	Primary Objective:
4	Create a local Obsidian-compatible vault that functions as a durable context layer for LLMs and agents. The output must be a working vault folder, not merely a report. The vault must be source-backed, resumable, validated by scripts, and organized around canonical knowledge rather than raw dumps.
5
6	Before doing substantive work, confirm:
7	1. The current working directory.
8	2. The intended output root directory.
9	3. The available source locations and connectors.
10	4. The connected account/workspace/organization for every external connector.
11
12	If the user has not provided an output root, create the compiled vault under the current working directory:
13
14	Compiled-Vaults/compiled-vault-brain-YYYY-MM-DD/
15
16	Do not write the final vault to an ambiguous location.
17
18	Core Memory Architecture:
19	The vault must distinguish between declarative memory, procedural memory, source traces, context packs, and runtime tools/connectors.
20
21	Required First Phase: Orientation:
22	Do not author the final vault immediately. First inspect available vaults, notes, documents, repos, generated examples, and connected tools.
23
24	Produce:
25
26	Reports/ORIENTATION-REPORT.md
27
28	Include confirmed working directory, output path, source inventory, connector inventory, account/workspace verification results, initial high-signal entities, proposed ingestion plan, and blockers.
29
30	Hard Checkpoint 1:
31	After ORIENTATION-REPORT.md is created, pause and ask the user whether to proceed. Do not begin broad ingestion until the user approves.
32
33	Connector Verification Gate:
34	Before ingesting from any external connector, retrieve and document the actual connected account/workspace/organization.
35
36	For each connector, record:
37	- Connector name.
38	- Account email or user ID if available.
39	- Workspace, organization, tenant, or team ID if available.
40	- Verification method.
41	- Timestamp.
42	- Read/write capability observed.
43	- Whether it is approved for ingestion.
44
45	Write this to SOURCE-MANIFEST.md.
46
47	If a connector is attached to the wrong account, workspace, or organization, mark it blocked. Do not use it for ingestion. Never send emails, messages, calendar invites, refunds, billing updates, database writes, or any external mutation without explicit user approval.
48
49	Required State And Resumption Files:
50	Create and maintain:
51
52	INGESTION-LOG.md
53	state.json
54
55	The job must be resumable from these files. Before doing new work after any interruption, read both files and continue from recorded state.
56
57	state.json must include output_path, current_phase, completed_phases, sources_discovered, sources_ingested, connector_status, batches_completed, canonical_notes_created, validation_status, next_actions, and blockers.
58
59	Required Compiler Process:
60	Use this repeatable compiler-style process:
61
62	parse -> thread/group -> classify -> extract -> canonicalize -> rehydrate provenance -> author -> validate links -> critic/audit
63
64	When re-encountering an existing entity, update the existing canonical note and append new provenance. Do not create a duplicate note unless there is clear evidence that it is a distinct entity.
65
66	For large ingestion passes, use a small or cheap model for parse, classify, and extract where available. Reserve the strongest model for synthesis, authoring, critic review, and ambiguous entity resolution.
67
68	Required Deterministic Machinery:
69	Create a tools folder inside the compiled vault:
70
71	_tools/
72
73	Write and run scripts for:
74
75	1. Wikilink validation:
76	- Find all [[wikilinks]].
77	- Mirror Obsidian resolution semantics: resolve links by note filename anywhere in the vault, support paths where specified, account for aliases/headings/blocks where practical, and flag only links that would not resolve in Obsidian.
78	- Report unresolved and ambiguous links.
79
80	2. Slug/path validation:
81	- Detect empty filenames.
82	- Detect invalid paths such as People/.md, Projects/.md, Companies/.md.
83	- Detect duplicate canonical slugs.
84	- Detect unsafe filename characters.
85
86	3. Secret and credential scanning:
87	- Scan output files for API keys, access tokens, private keys, passwords, recovery codes, bearer tokens, OAuth tokens, database URLs, SSH keys, webhook secrets, signing secrets, session cookies, and credential-like environment variables.
88	- Redact before completion.
89
90	4. Source reference validation:
91	- Confirm every promoted canonical note has provenance.
92	- Confirm every source reference points to SOURCE-MANIFEST.md or Sources/.
93
94	5. Required artifact validation:
95	- Confirm all required folders and files exist.
96
97	6. Manifest generation or validation:
98	- Generate or verify SOURCE-MANIFEST.md from ingestion logs and connector checks.
99
100	Record script commands and results in VALIDATION-REPORT.md.
101
102	Redaction Policy:
103	Never copy secrets or credential values into the compiled vault. For ordinary personal information, avoid unnecessary raw dumps. Prefer summaries and provenance over copying long private messages. If sensitive personal, legal, medical, financial, or employment information is needed, summarize minimally and mark sensitivity in frontmatter.
104
105	Required Ingestion Plan:
106	Existing notes/vaults:
107	- Ingest useful Markdown broadly.
108	- Prioritize people, companies, projects, daily notes, meeting notes, tasks, reports, resources, decisions, procedures, glossaries, and planning files.
109	- Do not mutate the original vault unless explicitly approved.
110
111	Email, if verified:
112	- Smoke pass: 5-10 high-signal threads.
113	- Create sample notes, source traces, and validation outputs.
114	- Pause for user approval before broader ingestion.
115	- Recent broad pass: approximately 300-500 relevant threads from the last 60-120 days.
116	- Entity-targeted pass: 500-1,500 additional threads focused on high-signal entities.
117	- Deep historical pass: 2,000-5,000 additional threads only if explicitly approved.
118
119	Hard Checkpoint 2:
120	After the email or connector smoke pass, show:
121	- 3-5 sample canonical notes.
122	- 1 sample source trace.
123	- 1 sample context pack.
124	- Current SOURCE-MANIFEST.md.
125	- Current VALIDATION-REPORT.md.
126
127	Then ask the user whether to proceed to broader ingestion.
128
129	Required Vault Structure:
130	The compiled vault must contain:
131
132	People/
133	Companies/
134	Projects/
135	Products/
136	Topics/
137	Decisions/
138	Commitments/
139	Procedures/
140	Preferences/
141	Context Packs/
142	Sources/
143	Maps/
144	Reports/
145	_tools/
146	README.md
147	SOURCE-MANIFEST.md
148	VALIDATION-REPORT.md
149	COMPLETION-AUDIT.md
150	INGESTION-LOG.md
151	state.json
152
153	Note Format:
154	Use Obsidian-compatible Markdown with YAML frontmatter where useful, wikilinks, aliases, tags, source references, and clear provenance. Prefer durable canonical notes over shallow contact cards. Separate claims from source traces. Represent uncertainty explicitly. Do not invent facts.
155
156	Example People Note:
157	---
158	type: person
159	name: "Example Person"
160	aliases:
161	- "E. Person"
162	tags:
163	- person
164	source_status: source-backed
165	sensitivity: normal
166	last_verified: YYYY-MM-DD
167	---
168
169	# Example Person
170
171	## Summary
172	Example Person is a [role/relationship] connected to [[Example Company]] and [[Example Project]].
173
174	## Relationship To User
175	- Relationship: [customer / collaborator / investor / vendor / teammate / friend / unknown]
176	- Context: [brief source-backed explanation]
177	- Confidence: high / medium / low
178
179	## Known Preferences Or Patterns
180	- [Observed preference or pattern]
181	Source: [[Sources/example-source-id]]
182
183	## Related Projects
184	- [[Example Project]]
185
186	## Open Loops
187	- [[Commitments/example-commitment]]
188
189	## Provenance
190	- [[Sources/example-source-id]] — [brief description, date]
191
192	Example Decision Note:
193	---
194	type: decision
195	decision_date: YYYY-MM-DD
196	status: active
197	tags:
198	- decision
199	source_status: source-backed
200	confidence: medium
201	---
202
203	# Decision: Example Decision Title
204
205	## Decision
206	[State the decision clearly.]
207
208	## Rationale
209	[Explain why this decision was made, based only on available evidence.]
210
211	## Alternatives Considered
212	- [Alternative 1]
213	- [Alternative 2]
214
215	## Consequences
216	- [Expected or observed consequence]
217	- [Operational implication]
218
219	## Related
220	- [[Example Project]]
221	- [[Example Person]]
222
223	## Provenance
224	- [[Sources/example-source-id]] — [what this source proves]
225
226	Example Context Pack:
227	---
228	type: context-pack
229	scope: "Example bounded task"
230	tags:
231	- context-pack
232	last_updated: YYYY-MM-DD
233	---
234
235	# Context Pack: Example Task
236
237	## Use This When
238	Use this context pack when an agent needs to [specific task].
239
240	## Essential Context
241	- [[Example Project]] is the main project.
242	- [[Example Person]] is the key stakeholder.
243	- [[Decisions/example-decision]] governs the current approach.
244
245	## Current State
246	[Concise source-backed summary.]
247
248	## Relevant Procedures
249	- [[Procedures/example-procedure]]
250
251	## Open Loops
252	- [[Commitments/example-commitment]]
253
254	## Sources
255	- [[Sources/example-source-id]]
256
257	## Agent Instructions
258	- Do not assume facts not present in linked notes.
259	- Check SOURCE-MANIFEST.md for source freshness.
260	- If acting externally, ask for explicit approval first.
261
262	Required Content:
263	Create notes for important people, companies, projects, products, systems, topics, decisions, commitments, procedures, preferences, context packs, maps/indexes, and source traces.
264
265	Validation Hard Gates:
266	Do not declare this goal complete unless all of these pass:
267
268	- 0 placeholder source references.
269	- 0 broken internal wikilinks under Obsidian-style resolution.
270	- 0 empty slugs.
271	- 0 invalid paths such as People/.md, Projects/.md, Companies/.md.
272	- 0 copied secrets, API keys, tokens, passwords, private keys, recovery codes, or credential values.
273	- Every promoted canonical note has provenance.
274	- Every connector used is listed in SOURCE-MANIFEST.md.
275	- Every connector account/workspace/organization status is documented.
276	- README.md explains how the vault is organized and how agents should use it.
277	- VALIDATION-REPORT.md lists scripts run, commands used, pass/fail status, and remaining issues.
278	- COMPLETION-AUDIT.md maps every explicit requirement in this prompt to evidence in the output.
279	- INGESTION-LOG.md and state.json are current.
280	- The final vault opens as a normal Obsidian-compatible folder.
281
282	Completion Audit:
283	COMPLETION-AUDIT.md must map every major requirement in this prompt to:
284	- Status: pass / partial / blocked / fail.
285	- Evidence path.
286	- Notes created or scripts run.
287	- Remaining limitation, if any.
288
289	Execution Rules:
290	Work in phases. Continue across resumptions by reading INGESTION-LOG.md and state.json first. Avoid repeating completed work. Keep concise progress updates. Use source-of-truth checks over speculation. Prefer actual filesystem, connector, repo, and document evidence to memory. Preserve user-owned source files. Do not make destructive changes. Do not perform external writes or mutations without explicit approval.
291
292	Stop Conditions:
293	Pause and ask the user before:
294	- Broad connector ingestion beyond smoke pass.
295	- Deep historical email/chat ingestion.
296	- Any external write or mutation.
297	- Any destructive local file operation.
298	- Any ingestion likely to be expensive, slow, or privacy-sensitive.
299
300	Quality Bar:
301	The compiled vault should not be a dump. It should be a structured, source-backed operating memory that helps future agents understand who the user is, what they are working on, who and what matters, what has been decided, what remains open, what procedures work, what preferences matter, and which sources justify those conclusions.
302
303	Completion Rule:
304	Do not declare the goal complete until the compiled vault exists, validation scripts pass, and SOURCE-MANIFEST.md, VALIDATION-REPORT.md, COMPLETION-AUDIT.md, INGESTION-LOG.md, and state.json all exist and prove the requirements were satisfied. If blocked, continue with available sources and document the blocker precisely rather than silently degrading the output.

What this prompt is good for

Landing pages

Create stunning product landing pages

SaaS websites

Modern SaaS marketing websites

Product launches

Launch pages for new products

Startup websites

Complete websites for early-stage startups

Similar prompts

View all

Original X post