What does this skill do, and when should you use it?
This skill analyzes taxable accounts for positions with unrealized losses and prioritizes candidates by potential tax benefit, holding period, and loss percentage. It estimates the effects of short- and long-term losses, including the $3,000 net-loss deduction limit and carryforwards. It proposes replacement securities with similar exposure while checking wash-sale risks across household accounts. Deliverables include an opportunity list, trade execution sheet, wash-sale calendar, tax-savings summary, and replacement rationale; qualified professionals should review the output.
Uses user-provided taxable-account holdings, cost basis, current value, holding period, realized gains and losses, and carryforwards; identifies and ranks unrealized-loss candidates; calculates a target harvesting amount and estimated tax savings; suggests replacements that are not substantially identical; checks 30-day lookback and forward windows, household taxable and retirement accounts, spouse accounts, and DRIPs; documents each trade, wash-sale window, risk, estimated loss, and replacement; produces an opportunity list, execution sheet, tracking calendar, summary, and rationale.
- A wealth advisor screens a client's taxable portfolio for year-end harvesting opportunities.
- An investor with realized gains evaluates how short- and long-term losses could offset them.
- An advisor checks household accounts for recent purchases, planned repurchases, and DRIPs before selling a security.
- An investment team selects replacement securities while preserving asset-class, sector, or geographic exposure.
- A tax or investment professional documents cost-basis updates, wash-sale windows, and tax-reporting records after execution.
How do you install this skill?
- The content assumes U.S. tax rules and must be reviewed by a qualified tax professional before use in China or another jurisdiction.
- Substantially-identical security treatment, retirement-account effects, tax rates, and carryforward rules require case-by-case confirmation and should not be used to execute trades directly.
- Human approval, permission boundaries, data protection, and failure handling are not specified; treat outputs as drafts requiring review.
In Cowork, open Settings → Plugins → Add plugin, paste https://github.com/anthropics/financial-services, and select the wealth-management vertical from the marketplace list. The README also supports uploading a zip of a directory under plugins/; it does not provide a dedicated installation command for this individual skill.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/anthropics/financial-services.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/plugins/vertical-plugins/wealth-management/skills/tax-loss-harvesting" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- year-end tax planning.
After installation, enter a trigger such as “tax-loss harvesting,” “TLH,” “harvest losses,” “tax losses,” “unrealized losses,” or “year-end tax planning.” Provide the relevant account and holdings data, then request the candidate list, replacement securities, wash-sale windows, and execution plan. Treat the output as draft work product, not tax, legal, accounting, or investment advice, and obtain qualified professional review before acting.
What are this skill's strengths and limitations?
- Covers candidate identification, tax budgeting, replacement selection, wash-sale checks, execution planning, and post-harvest tracking.
- Explicitly checks all household accounts, 30-day windows, and dividend reinvestment plans.
- Accounts for short- and long-term losses, the $3,000 ordinary-income deduction limit, and carryforwards.
- Produces concrete tables, a calendar, and replacement-security rationale.
- SKILL.md provides no data connector, automated account access, or trade-execution capability.
- No test suite, sample inputs, or platform validation record is provided.
- Replacement examples and tax rules require case-specific verification, and substantially-identical determinations can be complex.
- Transaction costs, tracking error, and future cost-basis effects may make some losses uneconomical to harvest.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Tax-Loss Harvesting Advisor this page ✓ Anthropic · Official | 49 · Use with care | ★ 40k | 1mo ago | Apache-2.0 |
| Portfolio AI Readiness Scan ✓ Anthropic · Official | 46 · Use with care | ★ 40k | 1mo ago | Apache-2.0 |
| Portfolio Rebalancing Advisor ✓ Anthropic · Official | 47 · Use with care | ★ 40k | 1mo ago | Apache-2.0 |
| Alpha Insights BizAdvisor | 59 · Recommended | ★ 68 | 3mo ago | MIT |
| Client Performance Report ✓ Anthropic · Official | 43 · Not recommended | ★ 40k | 1mo ago | Apache-2.0 |
How did FollowSkills review this skill?
The skill instructs review of all household accounts, DRIPs, and wash-sale windows, and frames trades as an execution plan, showing some awareness of data scope and external effects. It does not specify least privilege, data flows, sensitive tax-data handling, human confirmation, rollback, or authoritative tax sources, so points are deducted.
The workflow is organized across candidate identification, gain/loss budgeting, replacement securities, wash-sale checks, and tracking, making the key path understandable. It lacks input validation, abnormal-input handling, failure feedback, data-availability assumptions, and skill-specific tests; under static calibration it remains below 10.
The description, trigger phrases, and target scenario are clear for tax-loss harvesting in taxable accounts. However, non-fit cases, jurisdiction, account-data formats, and Chinese-language support are unspecified, while the tax rules are plainly U.S.-oriented, so points are deducted.
The documentation is readable and structured, with tables and important notes; the repository also supplies an Apache-2.0 license, official provenance, maintenance scripts, and CI signals. The skill itself lacks installation/dependency notes, versioning, changelog, maintenance ownership, FAQs, and troubleshooting guidance, so points are deducted.
The workflow covers opportunity screening, tax estimation, replacement securities, wash-sale management, an execution plan, and post-harvest tracking, giving the core task a plausible shape. However, no representative output is verified, Excel delivery dependencies are unstated, and substantial professional tax review remains necessary; the static ceiling supports at most 7.
Repository-wide plugin validation, secret scanning, and read-only CI evidence provide limited auditability. There are no tests for this skill, third-party tax citations, executed examples, or independently reproducible results, so only 4 points are awarded.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →