Obsidian Brokerage Activity Manager

Default read scope stays inside:

• 20 Resources/Investments/Brokerage Activity/**/*.md
• 20 Resources/Investments/Brokerage Activity/Brokerage Activity.base
• 20 Resources/Investments/Brokerage Assets/*.md
• 20 Resources/Investments/Brokerage Assets/Brokerage Assets.base
• brokerage export files explicitly named in the request
• any inbox capture note the user points to for context

3. Validate Existing Ledger Notes

uvx --from python --with pydantic --with pyyaml python \
  .skills/obsidian-brokerage-activity-manager/scripts/validate_brokerage_activity.py \
  --glob "20 Resources/Investments/Brokerage Activity/**/*.md"

Review invalid frontmatter and duplicate source_signature collisions before importing anything new.

Validate the derived asset surface too:

uvx --from python --with pydantic --with pyyaml python \
  .skills/obsidian-brokerage-activity-manager/scripts/validate_brokerage_assets.py \
  --glob "20 Resources/Investments/Brokerage Assets/*.md"

4. Dry-Run the Import

uvx --from python --with polars --with pydantic --with pyyaml python \
  .skills/obsidian-brokerage-activity-manager/scripts/sync_brokerage_activity.py \
  --input "brokerage-export.csv" \
  --provider auto \
  --mode check

The sync script:

• detects known providers from headers
• normalizes provider rows into one typed activity schema
• merges duplicate rows through source_signature instead of creating duplicate notes
• preserves stable note paths so reruns are idempotent
• derives one asset note per ticker symbol and keeps it aligned with the current activity surface
• warns when a date-windowed activity workbook has no earlier ledger history, because derived current holdings may omit pre-window positions
• supports future providers through canonical column names or an explicit --column-map

5. Apply the Import

uvx --from python --with polars --with pydantic --with pyyaml python \
  .skills/obsidian-brokerage-activity-manager/scripts/sync_brokerage_activity.py \
  --input "brokerage-export.csv" \
  --provider auto \
  --mode fix

When provider detection is too weak, pass an explicit mapping:

uvx --from python --with polars --with pydantic --with pyyaml python \
  .skills/obsidian-brokerage-activity-manager/scripts/sync_brokerage_activity.py \
  --input "future-broker.csv" \
  --provider generic_csv \
  --column-map '{"activity_date":"Executed At","activity_type":"Description","symbol":"Ticker","amount":"Net Amount","price":"Unit Price","quantity":"Units","status":"Status","currency":"Currency","brokerage":"Brokerage"}' \
  --mode fix

6. Render the Bases

uvx --from python --with pydantic --with pyyaml python \
  .skills/obsidian-brokerage-activity-manager/scripts/render_brokerage_activity_base.py \
  --output "20 Resources/Investments/Brokerage Activity/Brokerage Activity.base"

uvx --from python --with pydantic --with pyyaml python \
  .skills/obsidian-brokerage-activity-manager/scripts/render_brokerage_assets_base.py \
  --output "20 Resources/Investments/Brokerage Assets/Brokerage Assets.base"

The rendered Bases surface:

• an activity ledger, trade flow, distributions, cash movements, and review queue
• an asset registry keyed by ticker symbol, plus income, trading, and review views

7. Re-Validate and Verify

uvx --from python --with pydantic --with pyyaml python \
  .skills/obsidian-brokerage-activity-manager/scripts/validate_brokerage_activity.py \
  --glob "20 Resources/Investments/Brokerage Activity/**/*.md"

uvx --from python --with pydantic --with pyyaml python \
  .skills/obsidian-brokerage-activity-manager/scripts/validate_brokerage_assets.py \
  --glob "20 Resources/Investments/Brokerage Assets/*.md"

obsidian read path="20 Resources/Investments/Brokerage Activity/Brokerage Activity.base"
obsidian read path="20 Resources/Investments/Brokerage Assets/Brokerage Assets.base"
obsidian search query="brokerage_activity_kind" limit=20 total
obsidian search query="brokerage_asset_kind" limit=20 total
obsidian unresolved total

Treat import warnings as workflow gates. If a date-window warning says the ledger has no history before the workbook start date, do not trust downstream current holdings as complete until you either import older history or reconcile against a current valuation/holdings snapshot.

Kind Rules

Use references/brokerage-activity-schema.md for the canonical note contract.

• trade_buy

  • Outflow activity for executed buys.
  • Keep quantity and unit_price when the export provides them.

• trade_sell

  • Inflow activity for executed sells.
  • Preserve fee_amount when brokerage is explicit.

• distribution

  • Cash distributions or dividends credited to the account.
  • Usually pair with a same-day distribution_reinvestment in DRP workflows.

• distribution_reinvestment

  • Outflow generated by a distribution reinvestment plan.
  • Keep share quantity and unit price when supplied.

• cash_deposit and cash_withdrawal

  • Pure account-level cash movements without a security leg.

• fee, tax, cash_interest

  • Cash-account adjustments that should remain queryable in the Base.

• fx, corporate_action, adjustment

  • Reserve for events that are real but not cleanly expressible as the simpler kinds.
  • Default review_status should remain needs_review until the mapping is explicit.

Use references/brokerage-asset-schema.md for the asset contract.

• listed_security

  • Default asset note kind for tickered securities in brokerage exports.
  • Path identity is the ticker symbol, so re-imports update AAA.md, NDQ.md, and similar notes instead of creating duplicates.

• fund, crypto, fx_pair, other, unknown

  • Reserve for future providers or richer instrument classification once the export surface supports it.

Guardrails

• Treat brokerage_activity_kind as the supertag and schema selector for every generated note.
• Treat brokerage_asset_kind as the supertag and schema selector for every generated asset note.
• Prefer one canonical note per economic event, not one note per import file.
• Prefer one canonical asset note per ticker symbol.
• Deduplicate through stable source_signature values and merge source_row_hashes instead of duplicating notes.
• Keep asset notes derived from the current activity set; do not accumulate stale asset counters across re-syncs.
• Preserve non-owned frontmatter fields and manual tags on existing ledger notes when resyncing.
• Do not infer missing quantity, price, or ticker values when the export does not provide them.
• Prefer explicit --column-map input over brittle fuzzy parsing for unknown providers.
• Treat date-windowed activity workbooks as incremental ledger slices, not proof of complete current holdings, unless pre-window history is already in the vault or a current valuation reconciles the result.
• Keep each Base filtered to its typed folder and note surface only.

References

• references/brokerage-activity-schema.md — canonical schema, note path rules, and _kind taxonomy
• references/brokerage-asset-schema.md — ticker-keyed asset schema and aggregation rules
• references/provider-mapping.md — provider detection, column aliases, and future-provider extension rules
• scripts/brokerage_models.py — source of truth for the Pydantic v2 schema