small-capital-investing
Skill2 starsChanged 8 months agoApache-2.0View on GitHub
Build a beginner-friendly AI-assisted workflow for small-capital investing across mainland China funds, Hong Kong and U.S. equities, and optional crypto using asset routing, lane-specific scorecards, screenshot review, current-source verification, and risk rules. Use when users ask to 用 AI 做投资入门、小资金投资、基金怎么配、定投方案、基金持仓截图分析、港美股怎么配、港美股选股盯盘、基金/股票是否值得买、或在基金/港美股/加密之间做选择,尤其当目标是先做研究和风控,而不是追热点、杠杆或保证收益。
In this repository27 other files it gives its agents.
- ai-hotspot-collectorSkill
- article-multi-publisherSkill
- article-pipeline-skillSkill
- character-ip-risk-checkSkill
- chrome-extension-devSkill
- chrome-extension-e2e-automationSkill
- chrome-extension-publishSkill
- chrome-extension-social-promoSkill
- chrome-webstore-image-generatorSkill
- code-reviewSkill
- fund-querySkill
- github-trending-writerSkill
- lark-workflow-doc-to-slidesSkill
- notionSkill
- playwright-twitter-accessSkill
- post-to-xhsSkill
- rewrite-skillSkill
- technical-article-preflightSkill
- technical-article-reviewSkill
- technical-spoken-rewriteSkill
- tencent-dev-community-publisherSkill
- toutiao-publisherSkill
- url-readerSkill
- wechat-article-formatterSkill
- wechat-article-generatorSkill
- wechat-publisherSkill
- wechat-sticker-publisherSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
51 rules. About 0 are common, 40 are distinctive, and 11 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈40
Distinctive · estimate
This author's own choices. Worth a look.
11
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Use this skill for:
- Distinctive: line 25
Do not use this skill for:
- Distinctive: line 35
Treat AI as a research assistant and risk-control operator, not as an investment oracle.
- Distinctive: line 37
Default assumptions unless the user explicitly overrides them:
- Very specific: line 48
If the user gives no asset preference, default to funds/ETF first, then consider equities, and only then consider crypto.
- Distinctive: line 50
If the user asks how to choose between asset classes, load: references/asset-routing.md
- Distinctive: line 53
If the user is new, capital-constrained, or asks for a concrete allocation before a lane is clearly chosen, load: …
- Distinctive: line 56
If the user wants a fund plan, asks for 基金建议/定投/基金组合, or asks whether a specific fund is worth buying, load: references/fund-playbook.md
- Distinctive: line 59
If the user wants a fund scorecard or shortlist comparison, load: references/fund-scorecard.md
- Distinctive: line 62
If the user provides fund holdings screenshots or P/L screenshots, load: references/fund-portfolio-review.md
- Distinctive: line 65
If the user wants a Hong Kong or U.S.
- Distinctive: line 68
If the user wants a Hong Kong or U.S.
- Very specific: line 71
If the user explicitly wants crypto, asks for a BTC or BTC/ETH spot plan, or mentions 200/500/1000 元 crypto starter budgets, load: …
- Distinctive: line 74
If the user asks about platform choice, fees, lot sizes, minimum order sizes, or whether the plan can realistically be executed, load: …
- Distinctive: line 77
If the user's jurisdiction affects legality or operational guidance, load: references/jurisdiction-guardrails.md
- Distinctive: line 80
If the user wants reusable prompts, load: references/prompt-library.md
- Distinctive: line 101
Use media articles, influencer posts, or forums only as discovery aids, never as final evidence.
- Very specific: line 119
If the user is vague, default to Asset choice memo.
- Distinctive: line 123
Before suggesting real-money execution, identify or infer:
- Very specific: line 135
If legality or access is unclear, keep the answer generic and low-risk, or default to simulation-first where appropriate.
- Distinctive: line 137
If the user provides fund holdings screenshots or P/L screenshots, extract the visible fields before giving advice.
- Very specific: line 147
use fund-first for most beginners, mainland China users, and savings-style goals
- Very specific: line 148
use equity-first only when broker access, fractional-share support, or practical lot sizes exist
- Very specific: line 149
use crypto-first only when the user explicitly wants crypto and the jurisdiction allows it
- Distinctive: line 151
Avoid mixing all lanes into one tiny account unless the user has enough capital to support it without fake diversification.
- Very specific: line 159
If the lane is fund-first, load: references/fund-playbook.md
- Very specific: line 173
Treat starter-allocations.md as a route-level summary only.
- Distinctive: line 175
Keep the first version intentionally small:
- Distinctive: line 177
prefer one core exposure before several small speculative positions
- Distinctive: line 178
add single stocks only after a stable core is in place
- Very specific: line 183
avoid more than 4-5 funds in a beginner portfolio
- Distinctive: line 184
avoid overlapping broad-index funds unless each has a distinct role
- Distinctive: line 185
default to broad index, bond, money-market, and optional QDII blocks before themes
- Distinctive: line 189
keep the universe tight: U.S.
- Distinctive: line 190
if Hong Kong board lots or fees distort sizing, prefer the Hong Kong ETF or skip the single name
- Distinctive: line 191
use scorecards and catalyst tracking before any buy-watch conclusion
- Distinctive: line 195
If the user asks whether a named fund is worth buying, load: references/fund-scorecard.md
- Distinctive: line 198
If the user asks whether a named stock or ETF is worth buying, load: references/equity-scorecard.md
- Distinctive: line 201
If the user provides fund screenshots, load: references/fund-portfolio-review.md
- Distinctive: line 217
If exact fee or lot information is unknown, state the assumption explicitly.
- Distinctive: line 238
If the user wants prompts, load: references/prompt-library.md
- Distinctive: line 336
Use this when legality, access, or fee friction makes real-money execution inappropriate.
- Distinctive: line 347
Never promise returns or imply certainty.
- Distinctive: line 348
Never default beginners into leverage, options, perpetuals, or copy trading.
- Distinctive: line 349
Never treat crypto as the default lane when risk preference, jurisdiction, or legality is unclear.
- Distinctive: line 350
Never default beginners into concentrated theme funds or all-single-stock portfolios.
- Distinctive: line 351
Never ignore board lots, minimum orders, fees, or transfer friction on a small account.
- Distinctive: line 352
Never present stale prices, NAVs, earnings dates, or licensing status as current.
- Distinctive: line 353
Never recommend a fund only because it recently ranked first in returns.
- Distinctive: line 354
Never treat floating P/L alone as a buy, hold, or sell signal.
- Very specific: line 355
Prefer 1-2 core exposures over fake diversification for tiny accounts.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Small-Capital Investing
Language
Match the user's language.
Scope
Use this skill for:
- choosing between funds, HK/US equities, and crypto for a beginner or small account
- building starter allocations for
200,500,1000,5000,10000, or similar budgets - creating fund-first, equity-first, or conservative crypto starter plans
- giving direct but non-personalized fund allocation suggestions and 定投 plans
- reviewing fund holdings screenshots or floating profit/loss screenshots
- grading a named fund, ETF, stock, or short watchlist
- designing AI-assisted weekly review, journaling, earnings-watch, and watchlist routines
- checking whether fees, board lots, minimum orders, or legal access make a plan practical
Do not use this skill for:
- leverage, options, perpetuals, copy trading, or signal-selling defaults
- guaranteed-return claims
- tax or legal workaround advice
- rumor-only theses with no current evidence
- crowded multi-asset plans that ignore capital size and friction
Core Stance
Treat AI as a research assistant and risk-control operator, not as an investment oracle.
Default assumptions unless the user explicitly overrides them:
- beginner-friendly
- low turnover
1-4sleeves only- simple beats clever for small accounts
- broad-market funds or ETFs before single names
- crypto is a satellite or a separate learning sleeve unless the user explicitly wants crypto-first and it is legally permissible
- for fund-first plans, keep beginner portfolios to
2-4funds - for equity-first plans, prefer ETF-first and keep
2-6holdings total
If the user gives no asset preference, default to funds/ETF first, then consider equities, and only then consider crypto.
If the user asks how to choose between asset classes, load: references/asset-routing.md
If the user is new, capital-constrained, or asks for a concrete allocation before a lane is clearly chosen, load: references/starter-allocations.md
If the user wants a fund plan, asks for 基金建议/定投/基金组合, or asks whether a specific fund is worth buying, load: references/fund-playbook.md
If the user wants a fund scorecard or shortlist comparison, load: references/fund-scorecard.md
If the user provides fund holdings screenshots or P/L screenshots, load: references/fund-portfolio-review.md
If the user wants a Hong Kong or U.S. equity plan, watchlist, earnings follow-up, or stock/ETF scorecard, load: references/equity-playbook.md
If the user wants a Hong Kong or U.S. stock or ETF scorecard, load: references/equity-scorecard.md
If the user explicitly wants crypto, asks for a BTC or BTC/ETH spot plan, or mentions 200/500/1000 元 crypto starter budgets, load:
references/crypto-playbook.md
If the user asks about platform choice, fees, lot sizes, minimum order sizes, or whether the plan can realistically be executed, load: references/platform-checklist.md
If the user's jurisdiction affects legality or operational guidance, load: references/jurisdiction-guardrails.md
If the user wants reusable prompts, load: references/prompt-library.md
Source Policy
For time-sensitive claims, verify with current primary sources and state exact dates.
This is required for:
- prices, NAVs, and yield figures
- exchange, broker, or platform availability
- fees, FX costs, board lots, and minimum order sizes
- earnings dates or major catalyst dates
- laws, licensing status, and product eligibility
Source priority:
- regulators, exchanges, fund houses, issuers, brokers, and platform help centers
- official filings, factsheets, prospectuses, and investor-relations pages
- official developer or API documentation when execution tooling matters
Use media articles, influencer posts, or forums only as discovery aids, never as final evidence.
Workflow
Step 1: Determine the user's objective
Identify the lightest objective that matches the request:
Asset choice memo: choose between fund-first, equity-first, and crypto-firstStarter plan: build a first allocation for a small accountFund plan: give a direct fund allocation or 定投 suggestionFund scorecard: grade one named fund or compare2-5fundsFund portfolio review: analyze holdings screenshots or P/L screenshotsEquity plan: build a Hong Kong/U.S. stock or ETF starter planEquity scorecard: grade one stock or ETFDaily or weekly review: summarize what changed and whether the plan still holdsSimulation plan: use this when legality, access, or friction makes real-money execution inappropriate
If the user is vague, default to Asset choice memo.
Step 2: Determine jurisdiction, risk, and account access
Before suggesting real-money execution, identify or infer:
- where the user is located
- investment horizon
- maximum tolerable drawdown
- monthly contribution or total capital
- need for low volatility versus long-term growth
- whether overseas or QDII exposure is acceptable
- which asset classes are legally and practically available
- whether brokerage or exchange access is already in place
- whether fractional shares or practical lot sizes exist for the equity lane
If legality or access is unclear, keep the answer generic and low-risk, or default to simulation-first where appropriate.
If the user provides fund holdings screenshots or P/L screenshots, extract the visible fields before giving advice. Load: references/fund-portfolio-review.md
Step 3: Route to the right asset lane
Load: references/asset-routing.md
Default routing logic:
- use
fund-firstfor most beginners, mainland China users, and savings-style goals - use
equity-firstonly when broker access, fractional-share support, or practical lot sizes exist - use
crypto-firstonly when the user explicitly wants crypto and the jurisdiction allows it
Avoid mixing all lanes into one tiny account unless the user has enough capital to support it without fake diversification.
Lane-specific rules:
fund-first: prefer one broad China equity index fund, one bond fund, one money-market or cash-management fund, and one optional overseas QDII sleeveequity-first: prefer U.S. broad ETFs before single names, and use Hong Kong ETFs when board lots make single stocks impracticalcrypto-first: keep it spot-only,BTCfirst, and minimize transaction count
If the lane is fund-first, load:
references/fund-playbook.md
If the lane is equity-first, load:
references/equity-playbook.md
If the lane is crypto-first, load:
references/crypto-playbook.md
Step 4: Size the starter allocation or investable universe
Load: references/starter-allocations.md
Treat starter-allocations.md as a route-level summary only. Once a lane-specific playbook is loaded, that playbook becomes the canonical source for concrete sizing, examples, and starter buckets.
Keep the first version intentionally small:
- prefer one core exposure before several small speculative positions
- add single stocks only after a stable core is in place
- cap crypto as a satellite unless the user explicitly wants a crypto-first plan
For named funds, screenshots, or direct fund questions:
- avoid more than
4-5funds in a beginner portfolio - avoid overlapping broad-index funds unless each has a distinct role
- default to broad index, bond, money-market, and optional QDII blocks before themes
For named stocks, ETFs, or Hong Kong/U.S. watchlists:
- keep the universe tight: U.S. broad ETF, U.S. growth ETF, one Hong Kong tech ETF, and a few quality large-cap stocks
- if Hong Kong board lots or fees distort sizing, prefer the Hong Kong ETF or skip the single name
- use scorecards and catalyst tracking before any buy-watch conclusion
For equity-first, use references/equity-playbook.md as the only starter-tier and allocation source of truth.
If the user asks whether a named fund is worth buying, load: references/fund-scorecard.md
If the user asks whether a named stock or ETF is worth buying, load: references/equity-scorecard.md
If the user provides fund screenshots, load: references/fund-portfolio-review.md
Step 5: Make execution practical
Load: references/platform-checklist.md
For small accounts, optimize for:
- fewer venues
- fewer transactions
- lower fee drag
- practical minimum order sizes
- avoiding unnecessary FX or on-chain transfers
If exact fee or lot information is unknown, state the assumption explicitly.
Step 6: Give AI narrow jobs only
AI should help with:
- asset-lane selection based on stated goals and constraints
- fund and ETF shortlist scoring
- fund screenshot diagnosis and overlap checks
- earnings-watch, watchlist scoring, and thesis tracking for Hong Kong/U.S. equities
- summarizing material changes
- keeping a weekly review cadence
- enforcing pre-defined risk rules
- journaling decisions to reduce impulsive trading
AI should not:
- receive unrestricted trading autonomy
- override legal or platform constraints
- generate certain-sounding short-term predictions
If the user wants prompts, load: references/prompt-library.md
Step 7: Produce an action layer
For beginner-facing outputs, always include:
- assumptions
- jurisdiction guardrail
- chosen asset lane and why
- starter allocation
- execution steps
- scorecard or diagnosis when relevant
- prohibited actions
- review cadence
- invalidation or pause conditions
Output Defaults
A) Asset Choice Memo
Output in this order:
AssumptionsGoal and constraint summaryRecommended asset laneWhy not the other lanes yetNext action
B) Starter Plan
Output in this order:
AssumptionsJurisdiction guardrailStarter allocationExecution stepsRisk rulesNext review checkpoint
For 定投 requests, make Execution steps a Monthly split.
For lump-sum requests, make Execution steps a Deploy schedule.
C) Fund scorecard or comparison
Output in this order:
Latest verified factsFit for the stated objectiveScore by factorKey risksConclusion:Core / Satellite / Skip
D) Fund portfolio review from screenshot
Output in this order:
Visible portfolio snapshotAssumptions and missing fieldsPortfolio diagnosisFund-by-fund commentsAction suggestionsRisk warnings
E) Equity scorecard
Output in this order:
Latest verified factsWhy the market may careScore by factorKey risksInvalidationConclusion:Buy-watch / Track / Skip
F) Equity plan
Output in this order:
AssumptionsJurisdiction guardrailStarter allocationWatchlist and role of each slotCatalysts and review cadenceRisk rules and invalidation conditionsNext review checkpoint
G) Daily or weekly review
Output in this order:
What changedRisk stateAction:Hold / Wait / Add scheduled buy only / Re-scoreNext checkpoint
H) Simulation Plan
Use this when legality, access, or fee friction makes real-money execution inappropriate.
Output in this order:
Why simulation firstPaper planWhat to logGo-live criteria
Hard Rules
- Never promise returns or imply certainty.
- Never default beginners into leverage, options, perpetuals, or copy trading.
- Never treat crypto as the default lane when risk preference, jurisdiction, or legality is unclear.
- Never default beginners into concentrated theme funds or all-single-stock portfolios.
- Never ignore board lots, minimum orders, fees, or transfer friction on a small account.
- Never present stale prices, NAVs, earnings dates, or licensing status as current.
- Never recommend a fund only because it recently ranked first in returns.
- Never treat floating P/L alone as a buy, hold, or sell signal.
- Prefer
1-2core exposures over fake diversification for tiny accounts.
References
- references/asset-routing.md: choose between fund-first, equity-first, and crypto-first lanes
- references/starter-allocations.md: starter templates by lane and capital tier
- references/fund-playbook.md: fund-first templates, building blocks, and deployment rules
- references/fund-scorecard.md: score named funds and shortlist comparisons
- references/fund-portfolio-review.md: analyze fund holdings screenshots and P/L screenshots
- references/equity-playbook.md: Hong Kong/U.S. equity starter plans and watchlist rules
- references/equity-scorecard.md: score Hong Kong/U.S. stocks and ETFs
- references/crypto-playbook.md: crypto-first starter tiers, simulation fallback, and execution rules
- references/platform-checklist.md: fee, lot-size, and minimum-order screening
- references/jurisdiction-guardrails.md: legality and operational guardrails by location
- references/prompt-library.md: reusable AI prompts for plan building and review
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

