Files
goldbrain/goldbrain-memory.src/skills/goldbrain-memory/SKILL.md
T
2026-06-07 22:41:19 -05:00

425 lines
26 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
---
name: goldbrain-memory
description: Use the goldbrain Obsidian vault as Bryan's persistent memory across Claude/CoWork sessions. Use whenever the operator asks to remember, save, note, log, or capture anything durable — facts, preferences, decisions, schedule changes, commitments — or asks what Claude knows about them, what was discussed or decided before, to check their notes, load their memory/profile/context, add to their inbox, or to track a new project. Trigger even without memory phrasing when the request implies recalling or persisting state across sessions ("pick up where we left off", "anything on X before my meeting?"). Also use proactively at the start of substantive work sessions to load context and at the end to log outcomes. Do NOT use for Jason's ECHO vault, Obsidian/REST-API development questions, "memory" meaning RAM, timed reminders, email or local-file lookups, or generic second-brain advice.
---
# Goldbrain Memory
Use the **goldbrain** Obsidian vault as persistent memory. Read context accumulated across sessions; write things the operator asks to be remembered.
The primary operator is **Bryan Gilliom** (CEO, Message Point Media). Unless stated otherwise, "the operator" means Bryan. Jason is the vault's architect, not the day-to-day user. Write memory in third person ("Bryan prefers X", not "I prefer X") so the vault stays readable by humans and other agents.
## API Configuration
All calls use these constants — hardcoded for this personal plugin:
```
OBSIDIAN_BASE = https://goldbrainapi.mpm.to
OBSIDIAN_KEY = fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8
```
The endpoint has a **valid TLS certificate**, so `-k` is not needed (add it only if the cert ever changes to self-signed). Always pass the `Authorization: Bearer` header. Paths address the vault **at its root** (e.g. `/vault/_agent/...`).
**The goldbrain vault carries its own control docs.** Unlike Jason's ECHO vault, goldbrain keeps `CLAUDE.md` / `BOOTSTRAP.md` / `STRUCTURE.md` / `index.md` at the vault root, and those remain the canonical source of truth for the operating contract and structure. This plugin holds the day-to-day *procedure* (loading order, search-first, triage, scope switching, PATCH/append rules) and **defers** to the in-vault docs for the durable contract:
- Operating contract, principles, and safety rules: the vault's `CLAUDE.md`
- Bootstrapping an empty vault and repair: `references/bootstrap.md` (which defers to the vault's `BOOTSTRAP.md`)
- Vault layout and frontmatter conventions: `references/vault-layout.md` (mirrors the vault's `STRUCTURE.md`)
- Full API reference with every endpoint pattern and the memory routing map: `references/api-reference.md`
## Operating Contract & Safety
The vault is the **system of record** for long-term memory, not a scratchpad. Default to **additive updates, explicit status changes, and traceable summaries**. Keep agent-managed content (`agent_written: true` + `source_notes`) separable from human-authored content. Non-negotiable safety rules:
- Do not fabricate facts, relationships, or prior decisions.
- Do not mass-restructure the vault unless explicitly asked.
- Do not delete notes unless deletion is explicitly requested and clearly safe.
- Never store secrets or API keys inside a vault note.
Full contract lives in the vault's `CLAUDE.md`.
## When to Load Memory
Load at the start of any substantive conversation — anything beyond a single quick factual question. The signal: the operator is starting work, planning, asking for help with something that has state, or referencing prior discussions.
### Loading procedure
The cold-start reads are independent — **issue them in parallel** (one batch of 45 GETs), not sequentially. Parallel loading is ~3× faster wall-clock for the same call count.
| # | GET | Notes |
|---|-----|-------|
| 1 | `/vault/BOOTSTRAP.md` | The in-vault preflight/repair manifest. 404 → vault not set up; follow `references/bootstrap.md`. 200 → proceed (skim its checklist only if a structural path 404s later). |
| 2 | `/vault/_agent/memory/semantic/operator-preferences.md` | Bryan's profile |
| 3 | `/vault/_agent/context/current-context.md` | Active scope + Scope History |
| 4 | `/vault/_agent/sessions/` (listing) | Pick the ~5 most recent by reverse lex sort (filenames `YYYY-MM-DD-HHMM-<slug>.md`, so lex == chrono); only read the ones whose slugs look relevant |
| 5 | `/vault/journal/daily/YYYY-MM-DD.md` | Today's note; 404 is fine — it's created on first agent activity |
Do not read every session log — older sessions are reachable via `POST /search/simple/?query=...` when needed.
**If a specific project is in play**, follow up with a **search across all lifecycle subfolders** (`active/`, `incubating/`, `on-hold/`, `archived/`) — searching one folder at a time misses notes filed elsewhere. Search by **both the slug AND any human title** the operator used in this conversation:
```bash
# slug
curl -s -X POST -H "Authorization: Bearer fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8" \
"https://goldbrainapi.mpm.to/search/simple/?query=<slug>"
# human title (avoids missing notes filed under a different name)
curl -s -X POST -H "Authorization: Bearer fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8" \
"https://goldbrainapi.mpm.to/search/simple/?query=<human+title>"
```
Then read whichever match lives at `projects/<lifecycle>/<slug>.md`.
Do NOT narrate this loading. Reading memory is expected behavior.
## Inbox Triage
`inbox/captures/inbox.md` is the catch-all for "save this somewhere — figure out where later." Without triage it grows forever and durable facts get stranded there.
At the start of a substantive session, GET `inbox/captures/inbox.md`. If it contains lines older than ~7 days that haven't been routed elsewhere, surface them once:
> "Three captures from last week are still in the inbox — want to route them now or leave them?"
When routing accepted items, send each to its proper home:
- Preference / pattern → PATCH-append under `operator-preferences.md::Observations`
- Project idea → PUT `projects/incubating/<slug>.md`
- Durable fact → PUT `_agent/memory/semantic/<slug>.md`
- Person fact → PUT/PATCH `resources/people/<name>.md`
Then record the move in `inbox/processing-log/YYYY-MM-DD.md` (one line per item: `- <original line> → <destination path>`). Don't delete the original capture unless Bryan explicitly asks — the processing log is the audit trail.
## Project Lifecycle
Projects move through four folders under `projects/`. The folder name and the `status:` frontmatter field MUST agree — they are two views of the same state.
| Folder | `status:` | Meaning |
|--------|-----------|---------|
| `projects/incubating/` | `incubating` | Idea captured; not actively worked. Promote when work starts. |
| `projects/active/` | `active` | Current work. The default state for anything in motion. |
| `projects/on-hold/` | `on-hold` | Paused but still tracked. Resumable. |
| `projects/archived/` | `archived` | Done, abandoned, or rolled into something else. Not deleted. |
**Promotion / transition rule:** move the file to the new folder AND update `status:` in the same change. A file in `projects/active/` with `status: on-hold` is broken state — fix it when you see it.
**Searching:** `POST /search/simple/?query=<slug>` covers all four subfolders in one call. Always search before creating a new note (see the pre-write rule below).
## When to Write Memory
Write when the operator:
- States a fact, preference, or commitment worth keeping ("I prefer X", "we use uv not pip", "standup is Tuesday at 10")
- Makes a non-obvious decision worth recording
- Says "remember that", "save this", "log this", "add to memory", "note that"
- Finishes a meaningful working session future sessions should pick up
Write in third person about Bryan. Every note carries the canonical frontmatter (see below). Agent-generated notes set `agent_written: true`.
### Before you write — search first (MANDATORY for new notes)
**Before creating any new note at `projects/<lifecycle>/<slug>.md`, `_agent/memory/semantic/<slug>.md`, `resources/people/<name>.md`, or any other slug-addressed location, search the whole vault for that slug.** Listing a single folder (e.g. `projects/active/`) is NOT sufficient — a note with the same slug may exist in `projects/on-hold/`, `projects/incubating/`, `projects/archived/`, or under a different folder entirely.
```bash
curl -s -X POST \
-H "Authorization: Bearer fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8" \
"https://goldbrainapi.mpm.to/search/simple/?query=<slug>"
```
If a match is found:
- In a non-active project subfolder (`on-hold/`, `incubating/`, `archived/`): **promote/merge** — PUT the merged content to `projects/active/<slug>.md` with `status: active`, then DELETE the old location. Preserve the earliest `created:` date.
- In the same folder you intended to write: **update in place** (PATCH or merged PUT). Never silently overwrite — fold the existing content in first.
- Elsewhere (e.g. a stale duplicate under `resources/`): tell Bryan and ask which should be canonical before writing.
**Search both the slug AND any human title** the operator used (e.g. slug `goldbrain-memory` and title `goldbrain plugin`). Slug-only searches miss notes filed under a different naming scheme. Two cheap `POST /search/simple/?query=...` calls beat one expensive cleanup pass later.
Only after the search comes back empty (or you've decided to merge) is it safe to create a new note. This rule prevents the most common duplication bug: a note exists in `on-hold/` but the agent only checked `active/` and created a parallel record.
### Before you append — read first (idempotency)
POST appends to the end of a file. It is **not idempotent** — running the same write twice (network retry, replay, re-trigger) produces duplicate lines that grow files silently. Before any POST that adds an entry to:
- `inbox/captures/inbox.md`
- a daily note's `## Agent Log` section
- a `## Fact / Pattern` / `## Observations` / `## Log` heading
…GET the target file (or the heading, via `/heading/...`) and substring-search for the exact line you're about to write. If present, skip the POST. The extra GET is cheap and pays for itself within a few sessions.
This rule does **not** apply to PUT (which is fully replacing the file) or PATCH `replace` (which is overwriting a section by design).
### Append to a file (default — additive entries)
Write content to a temp file first to handle multi-line markdown cleanly, then POST:
```bash
cat > /tmp/obs_entry.md << 'OBSEOF'
- 2026-06-07: <your entry here>
OBSEOF
curl -s -X POST \
-H "Authorization: Bearer fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8" \
-H "Content-Type: text/markdown" \
--data-binary @/tmp/obs_entry.md \
"https://goldbrainapi.mpm.to/vault/inbox/captures/inbox.md"
```
POST appends to the end of a file (creating it if absent). Use it for inbox captures and log sections.
### Patch a specific heading (targeted update)
**Heading targets use the FULL heading path, `::`-delimited from the top-level heading** — a bare subheading name fails with `400 invalid-target` (errorCode 40080) and the write is lost. For example, `## Fact / Pattern` under the `# Operator Preferences` H1 is targeted as `Operator Preferences::Fact / Pattern`.
**Default: GET the document map first** (every first PATCH to a file in a session — cache the result mentally for subsequent PATCHes to the same file). This eliminates the most common failure mode of PATCH:
```bash
curl -s -H "Authorization: Bearer fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8" \
-H "Accept: application/vnd.olrapi.document-map+json" \
"https://goldbrainapi.mpm.to/vault/_agent/memory/semantic/operator-preferences.md"
```
Returns `{ "headings": [...], "blocks": [...], "frontmatterFields": [...] }`. Copy the heading string verbatim into `Target`. Only skip the doc-map GET if you wrote the file yourself in this session (you already know its structure).
Then PATCH:
```bash
cat > /tmp/obs_patch.md << 'OBSEOF'
Bryan prefers status updates that lead with the decision.
OBSEOF
curl -s -X PATCH \
-H "Authorization: Bearer fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8" \
-H "Operation: append" \
-H "Target-Type: heading" \
-H "Target: Operator Preferences::Fact / Pattern" \
-H "Content-Type: text/markdown" \
--data-binary @/tmp/obs_patch.md \
"https://goldbrainapi.mpm.to/vault/_agent/memory/semantic/operator-preferences.md"
```
Use `Operation: replace` to overwrite a section entirely (e.g. a project's `Project Name::Current status`).
#### Special characters in heading targets — em dashes, parentheses (KNOWN BUG)
The `Target` header is matched against the heading text only after the server URL-decodes it, and some characters break that match. A heading like `# EarthDnD — Personal Project` targeted literally returns `invalid-target` (40080), because raw non-ASCII bytes (the em dash `—`, U+2014) ride unreliably in an HTTP header and the matcher applies regex-special rules to characters like parentheses and brackets without escaping them. Three rules, in priority order:
1. **Prevent it — keep new headings ASCII (the durable fix).** When you create or rewrite a heading you may later PATCH, use a plain hyphen surrounded by spaces (` - `), never an em dash (`—`) or en dash (``), and avoid parentheses/brackets in heading text. Write `# EarthDnD - Personal Project`, not `# EarthDnD — Personal Project`. ASCII headings are always PATCH-safe. **Do not auto-format hyphens into em dashes in headings** — this is the root cause of the bug. (Em dashes in body prose are fine; the rule is about heading text that becomes a PATCH `Target`.)
2. **Target an existing special-char heading by percent-encoding the offending run.** Em dash `—` (U+2014) → `%E2%80%94`; en dash `` (U+2013) → `%E2%80%93`; spaces may stay literal or be `%20`. Reproduction case from the bug report, fixed:
```bash
# heading in file: # EarthDnD — Personal Project
curl -s -X PATCH \
-H "Authorization: Bearer fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8" \
-H "Operation: append" \
-H "Target-Type: heading" \
-H "Target: EarthDnD %E2%80%94 Personal Project::AI Platform Roles" \
-H "Content-Type: text/markdown" \
--data-binary @/tmp/obs_patch.md \
"https://goldbrainapi.mpm.to/vault/projects/active/earthdnd-personal.md"
```
3. **Fallback when encoding still fails (e.g. parentheses in the heading).** GET the whole file, edit the section in memory, and PUT the file back. This overwrites atomically rather than patching in place, so it carries a small race-condition risk on frequently-written files — use it sparingly and, where you control the heading, fix it to ASCII (rule 1) so the next write can PATCH cleanly.
See `references/api-reference.md` for the full encoding table.
### Bump `updated:` after meaningful changes
When a PATCH or PUT changes meaningful content (status update, decision recorded, current-status replacement, scope switch), also PATCH the frontmatter `updated:` field to today's date. This keeps stale-detection queries honest.
```bash
curl -s -X PATCH \
-H "Authorization: Bearer fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8" \
-H "Operation: replace" -H "Target-Type: frontmatter" -H "Target: updated" \
-H "Content-Type: application/json" --data '"2026-06-07"' \
"https://goldbrainapi.mpm.to/vault/projects/active/<slug>.md"
```
Skip the bump for **routine log appends** — adding an Agent Log line, an inbox capture, or a timestamped Observations bullet doesn't constitute a meaningful content change. Bump on substance, not on heartbeat.
### Create or overwrite a file (PUT)
```bash
cat > /tmp/obs_file.md << 'OBSEOF'
---
type: project
status: active
created: 2026-06-07
updated: 2026-06-07
tags: []
agent_written: true
source_notes: []
---
# Project Name
## Current status
...
## Related
- [[journal/daily/2026-06-07]]
OBSEOF
curl -s -X PUT \
-H "Authorization: Bearer fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8" \
-H "Content-Type: text/markdown" \
--data-binary @/tmp/obs_file.md \
"https://goldbrainapi.mpm.to/vault/projects/active/my-project.md"
```
The API creates intermediate directories automatically.
### Search the vault
```bash
curl -s -X POST \
-H "Authorization: Bearer fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8" \
"https://goldbrainapi.mpm.to/search/simple/?query=your+search+terms"
```
## Scope Switching (`current-context.md`)
`_agent/context/current-context.md` tracks a single active scope. The operator routinely shifts scope within a day (a project → MPM ops → something else).
When scope changes:
1. PATCH `prepend` a dated bullet to `## Scope History` capturing the **prior** scope (one line: `- 2026-06-07: <prior scope summary>`). If `## Scope History` doesn't exist yet, POST the heading first, same pattern as the daily-note Agent Log.
2. PATCH `replace` `## Scope` with the new scope.
3. PATCH the frontmatter `updated:` field.
This keeps a rolling trail of recent scopes in one file instead of spawning separate stash notes. Trim Scope History to the last ~10 entries when it grows past that.
## Vault Health (monthly)
On the first substantive session of a calendar month, run a quick health pass and write findings to `reviews/monthly/YYYY-MM-vault-health.md`. Don't auto-fix without asking.
Checks:
1. **Stale active projects** — for each note in `projects/active/`, check `updated:` >30 days. Likely belongs in `on-hold/`.
2. **Unprocessed inbox** — GET `inbox/captures/inbox.md`. List items older than 14 days that never moved through the triage protocol.
3. **Duplicate slugs across lifecycle folders** — any slug appearing in more than one of `active/`, `incubating/`, `on-hold/`, `archived/` is broken state.
4. **Broken-heading risk** — sample 23 frequently-PATCHed files; confirm `## Agent Log`, `## Scope`, `## Fact / Pattern`, `## Observations` headings still exist, and that no heading has acquired an em dash or parentheses that would break PATCH targeting.
The pass is cheap (a few searches + a directory listing) and pays for itself by catching drift before it requires a reorg.
## Daily Note — Agent Log
After substantive activity, write a one-line entry to today's daily note's `## Agent Log` heading. The daily-note **template** (`journal/templates/daily-note-template.md`) defines this heading, but ad-hoc daily notes created without the template don't have it — so PATCH can fail with `invalid-target` if the note exists but lacks the heading.
**Procedure (resilient):**
1. Try to GET `journal/daily/YYYY-MM-DD.md`.
2. If 404 — PUT a fresh daily note from the template, then PATCH.
3. If 200 but `## Agent Log` is missing — POST `\n\n## Agent Log\n` to the file to add the heading, then PATCH. **Detect the heading by grepping the raw markdown for an anchored `^## Agent Log` line — do NOT grep the document-map JSON, whose headings are full `::`-delimited paths (e.g. `"2026-06-07::Agent Log"`); a bare `"Agent Log"` match fails there and a duplicate heading gets appended.**
4. PATCH-append the entry under the target `<YYYY-MM-DD>::Agent Log` (the H1 of a daily note is the date, so that's the full target path).
```bash
DATE=$(date +%Y-%m-%d)
DAILY="https://goldbrainapi.mpm.to/vault/journal/daily/${DATE}.md"
AUTH="Authorization: Bearer fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8"
# 1+2: ensure the daily note exists (PUT from template if missing)
HTTP=$(curl -s -o /dev/null -w "%{http_code}" -H "$AUTH" "$DAILY")
if [ "$HTTP" = "404" ]; then
curl -s -H "$AUTH" "https://goldbrainapi.mpm.to/vault/journal/templates/daily-note-template.md" \
| sed "s/{{date:YYYY-MM-DD}}/${DATE}/g" \
| curl -s -X PUT -H "$AUTH" -H "Content-Type: text/markdown" --data-binary @- "$DAILY"
fi
# 3: ensure the `## Agent Log` heading exists.
# Grep the RAW markdown for an anchored heading line. Do NOT grep the document-map
# JSON: its headings are full "::"-delimited paths (e.g. "2026-06-07::Agent Log"), so a
# bare '"Agent Log"' match never matches and a DUPLICATE heading gets appended each time.
if ! curl -s -H "$AUTH" "$DAILY" | grep -q '^## Agent Log'; then
printf '\n\n## Agent Log\n' | curl -s -X POST -H "$AUTH" -H "Content-Type: text/markdown" --data-binary @- "$DAILY"
fi
# 4: PATCH-append the entry
cat > /tmp/agent_log.md << OBSEOF
- ${DATE}: <one-line entry here>
OBSEOF
curl -s -X PATCH -H "$AUTH" \
-H "Operation: append" \
-H "Target-Type: heading" \
-H "Target: ${DATE}::Agent Log" \
-H "Content-Type: text/markdown" \
--data-binary @/tmp/agent_log.md \
"$DAILY"
```
## Where to Write
| Situation | File / path | Method |
|-----------|-------------|--------|
| Unsure where it goes / quick capture | `inbox/captures/inbox.md` (date-prefixed line) | POST |
| Operator preference or durable fact | `_agent/memory/semantic/operator-preferences.md` (append under the right heading) | PATCH |
| Other durable facts / patterns | `_agent/memory/semantic/<slug>.md` | PUT |
| What happened (event record) | `_agent/memory/episodic/<slug>.md` | PUT |
| Short-lived working state | `_agent/memory/working/<slug>.md` | PUT |
| Task-scoped context / focus | `_agent/context/current-context.md` | PATCH / PUT |
| Working session ended with substance | `_agent/sessions/YYYY-MM-DD-HHMM-<slug>.md` | PUT |
| Long-running project state | `projects/<lifecycle>/<slug>.md` (see Project Lifecycle) | PUT + PATCH |
| Non-obvious decision (ADR) | `decisions/by-date/YYYY-MM-DD-<slug>.md` (see mirror note below) | PUT |
| Person context | `resources/people/<name>.md` | PUT / PATCH |
| Concept / reference note | `resources/concepts/` or `resources/references/` | PUT |
| Daily activity / Agent Log | `journal/daily/YYYY-MM-DD.md` — see **Daily Note — Agent Log** above | PATCH (with auto-create) |
**Decision mirrors:** mirror a project-relevant ADR as a `[[wikilink]]` under that project's `## Decisions` heading (PATCH), and optionally into `decisions/by-project/` per the vault's `STRUCTURE.md`. The by-date ADR is always the canonical record.
Never delete files unless Bryan explicitly asks. Memory is append-friendly; deletion is destructive.
## Session Logging
At the end of substantive conversations (ones that produced decisions, artifacts, or commitments), create a session log. Ask once if unsure: "Want me to log a session note for this?"
**Filename format is canonical: `_agent/sessions/YYYY-MM-DD-HHMM-<slug>.md`.** Always include the four-digit local-time HHMM component — it makes filenames lexically sort in true chronological order, which is what Step 3 of loading relies on. Older session logs without the HHMM part may exist; leave them alone, but every new one must use the full form.
See `references/session-log-template.md` for the body format.
```bash
cat > /tmp/obs_session.md << 'OBSEOF'
<session log content — see references/session-log-template.md>
OBSEOF
curl -s -X PUT \
-H "Authorization: Bearer fb72065a05fabb28ae87c45880cc3b7aba4fd3f58e70297934145cef974e8ed8" \
-H "Content-Type: text/markdown" \
--data-binary @/tmp/obs_session.md \
"https://goldbrainapi.mpm.to/vault/_agent/sessions/2026-06-07-1430-my-session.md"
```
Then add a one-line entry to today's daily note via the **Daily Note — Agent Log** procedure above.
## Vault Unreachable
If the API returns a connection error, timeout, or `502`, tell Bryan once that the memory vault is unreachable (a `502` usually means Obsidian/the REST plugin is not running on the backend), then proceed without memory. Do not retry repeatedly.
## Style Rules
- Write in third person about Bryan: "Bryan prefers X", not "I prefer X".
- Do not attribute Jason's design preferences to Bryan — capture Bryan's preferences only from observed interactions. Jason's preferences belong in the ECHO vault, Bryan's belong here.
- **Anchor relative dates on the conversation's `currentDate`** before writing. "Today" → `currentDate`. "Thursday" / "next week" → resolve to an absolute `YYYY-MM-DD`. Never guess from training-data knowledge of the current year.
- **Keep heading text ASCII** so it stays PATCH-targetable — plain hyphens, no em dashes or parentheses in headings (see the special-characters note under PATCH).
- Every memory file has canonical YAML frontmatter — see `references/vault-layout.md`.
- Set `agent_written: true` on agent-generated notes and list `source_notes` (plain relative paths, not links).
- **`created:` is the earliest known date the entity was tracked anywhere in the vault, not "today".** When merging notes (e.g. promoting `on-hold/` → `active/`), preserve the earliest `created:` and only bump `updated:`.
- **`source_notes` lists the note(s) that *triggered* or *supplied content for* this one** — e.g. the session log that produced a project update, or the daily note where a captured fact originated. It is a *backward* link to inputs. Forward links (this note → other notes it references) belong in the `## Related` section in the note body, never in frontmatter.
- **Never put `[[wikilinks]]` in frontmatter** — YAML parses them as nested lists and the links break in Obsidian's reading view. Put all cross-references in a `## Related` section in the note **body** as a bulleted list of `[[links]]`.
- Use Obsidian wiki links (`[[note name]]`) freely in the note **body** for cross-references.
- Keep entries short and focused. Fewer, sharper entries beat many noisy ones.
- About to write something large or sensitive? Show Bryan the content first and confirm.
## operator-preferences.md — Rules vs Observations
`_agent/memory/semantic/operator-preferences.md` separates two kinds of content:
- `## Fact / Pattern` — **promoted, deduped rules.** No date prefix. These are timeless: "Bryan prefers concise communication." Append here only when a rule is stable.
- `## Observations` — **timestamped raw observations.** Date-prefixed: `- 2026-06-07: Bryan chose X over Y because Z.` This is where new evidence goes by default.
During monthly Vault Health or when an observation stabilizes, promote it from `## Observations` into `## Fact / Pattern` (drop the date) and remove the duplicate from Observations. Trim Observations to the last ~30 entries when it grows past that — the rest live in session logs.
## What This Skill Does Not Do
- Does not replace reasoning. The vault is reference material — apply judgment.
- Does not auto-summarize the whole vault. Read targeted files, not everything.
- Does not store passwords or secrets, and never writes the API key into a vault note. If asked to "remember my password is X", decline and suggest a password manager.