Investors run due diligence on you. They ask for your data, metrics, traction. But who are they? Why don't they introduce themselves first?
Check them before you meet them. Before meeting, check the rating first.
When to Use
Use this skill when:
- The user needs to conduct a background check and rating on an investor before meeting them.
- The user has received a TS (Term Sheet) and needs to evaluate the quality of the fund.
- The user needs to determine whether a certain investor/fund is worth meeting.
- An FA recommended an investor, and the user needs to evaluate them.
- The user needs to compare multiple investors and decide whom to meet first.
- The user needs to screen for the most active investors in the AI sector or a specific round.
Do NOT use this skill if:
- The user needs to broadly research a list of potential investors → use
investor-researchinstead - The user needs to create a pitch deck → use
pitch-deck-creatorinstead - The user is seeking general business advice.
- The user needs to conduct due diligence on a company rather than an investor.
Typical triggers:
Direct triggers:
- "RateVC" / "check VC" / "Rate this VC" / "look up this investor"
- "help me look up this investor" / "rate this investor"
- "Check this investor" / "Score this investor" / "run a background check"
Name-based triggers (investor or fund name + evaluation intent):
- "Is [fund] worth meeting?" / "Is XX Fund reliable" / "How is XX Capital"
- "Is Zhu Xiaohu worth meeting" / "What do you know about [investor]?"
- "Does Sequoia still have money" / "Does [fund] still have money?"
- "Is he a GP or a VP" / "Is [person] a GP or just a VP?"
- "What has Shen Nanpeng been investing in recently" / "What has [fund] invested in recently?"
Decision-point triggers:
- "An investor contacted me, should I meet them" / "An investor reached out, should I take the meeting?"
- "This person asked to chat next week, is it worth going"
- "Should I fly to Beijing to meet this investor" / "Should I fly to SF to meet this investor?"
- "I received a TS, how is this fund" / "Got a term sheet from [fund], are they legit?"
- "An FA recommended several investors to me" / "An FA recommended this investor"
Comparison and screening triggers:
- "Which of these investors should I meet first" / "Who should I meet first?"
- "Help me list the most active investors in the AI sector"
- "Which investors are most worth meeting for a Series A"
- "Compare [fund A] vs [fund B]"
Risk check triggers:
- "Has this investor invested in my competitor"
- "Has he been in litigation with portfolio companies"
- "Does this fund still have money to invest"
Workflow
This workflow follows a six-step progression method, with 6 steps in total:
- Probe: Before starting, read all inputs completely, and confirm the goals and constraints.
- Constrain: Verify input completeness, set boundaries and delivery standards that cannot be downgraded, and switch channels when blocked.
- Evidence: Every number must come from the input, a specific source, or reproducible calculation, and supporting data must be collected.
- Execute: Call scripts/APIs/generate content, first give impact and conclusions, then give actions and necessary evidence.
- Verify: Verify with "possibly failing" actions—read back the output in a way different from the generation path.
- Deliver: Return the results and clean up temporary files.
Step 1: Probe
Identify the geographic region and the investor. Determine whether the investor is in China, the United States, or another region. If only an institution name is given, clarification is needed—the rating is forpersonrather than the institution.
Step 2: Constrain
Set boundaries and delivery standards that must not be downgraded:
- Never fabricate data—mark information that cannot be verified as "unverified" or downgrade the rating
- Do not fabricate deal counts, AUM figures, or positions
- Ratings are for individuals rather than institutions
- Do not engage in personal attacks—focus only on investment behavior data
- Do not attempt to access private databases, paid APIs without a subscription, or non-public financial data
- Language matching—always reply in the language used by the user
- The 1-4 scale and the 13 coverage rules are calibrated based on 120 real investor case studies, and the weights must not be deviated from without the user's explicit request
- "No Rating" is a valid output—not every investor can or should be rated
- When blocked, switch channels; do not downgrade the deliverable
Step 3: Evidence
Collect intelligence using public sources based on the region:
China:
- ITJuzi (itjuzi.com) — deals, industries, rounds
- Tianyancha / Qichacha — company registration, equity changes
- AMAC (Asset Management Association of China) — fund registration, AUM
- 36Kr / ChinaVenture — news, personnel changes
United States:
- Crunchbase — deals, portfolios, fund information
- SEC EDGAR — Form D, 13F filings
- LinkedIn — background, tenure
- PitchBook / Tracxn — in-depth deal data (if available)
⚠️ Data freshness rule: If the data is older than 90 days, mark it: "⚠️ Data last verified [date]"
Step 4: Execute
First give impact and conclusions, then give actions and necessary evidence. Score according to the following process:
Six-dimension scoring
| Dimension | Weight | 4 (Top) | 3 (Strong) | 2 (Adequate) | 1 (Base/Missing) |
|---|---|---|---|---|---|
| Dry Powder | 25% | Fund <2yr + deals in 3mo + >60% remaining | Active fund + deals in 6mo | Unclear / older fund | Depleted / no fund |
| Sector Fit | 25% | ≥3 deals in your space (12mo) + sector >50% of fund | ≥2 deals in space (12mo) | Related but imprecise | No match |
| Decision Power | 20% | Founding GP + solo authority + >10yr tenure | GP/Partner + can decide | Partner/MD + can recommend to IC | VP/Director/Associate |
| Activity | 15% | >15 deals/yr + deals in 3mo | 5-15/yr + deals in 6mo | <5/yr | Dormant >6mo |
| Lead Role | 10% | Leads >70% | Leads >50% | Mixed | Mostly follows / can't lead |
| Risk | 5% | Clean record | Minor controversy | Moderate risk | Hard red flag → veto |
13 coverage rules
Downgrade rules:
- Sector + Activity double-low: Both ≤1 → force downgrade one star.
- Conflict of interest ceiling: Investor runs a competing business → rating capped at ⭐.
- Stage mismatch: Investor only does B+ rounds, founder raising A → Sector -1.
- Power decay: Semi-retired / delegated daily decisions → Decision Power -1.
- CVC ceiling: CVC capped at ⭐⭐. Corporate approval + info leakage risk. Exception: parent has zero competitive overlap.
- Angel limitation: No fund vehicle → Lead Role = 1, Dry Powder ≤ 2.
Upgrade rules: 7. Sector focus bonus: Target sector >70% of fund portfolio → Sector +0.5 (capped at 4).
Exclusion rules (No Rating): 8. Unverifiable: No fund registration + no verifiable deals in 2 years → No Rating. 9. Departed: Retired / left firm / fund closed → No Rating. 10. Cooling period: Joined new fund <6 months ago → Pending, no rating.
Annotation rules: 11. Same fund, different people: Rating is for the person, not the firm. Must specify the individual. 12. FA / Banker tag: FAs and bankers don't invest. Dry Powder=1, Decision=1, Lead=1. Add ⚡tag with FA value explanation. 13. Fund structure is signal: E.g., Benchmark = all equal GPs → every partner is a core decision maker.
Michelin rating
| Rating | Score | Meaning | Action |
|---|---|---|---|
| ⭐⭐⭐ | ≥ 3.0 | Worth a Special Trip (worth a special trip) | Book a flight to meet them |
| ⭐⭐ | ≥ 2.0 | Worth Meeting In Person (worth meeting in person) | Same-city meeting or invite them over |
| ⭐ | ≥ 1.0 | Worth a Video Call (worth communicating online) | 30-60 min video call |
| No Rating | — | Don't Invest Time (not worth investing time for now) | Exclusion reason provided |
Stars = integer part of score. 3.5 = three stars, 2.3 = two stars, 1.7 = one star. Max 4, base 1.
Step 5: Verify
Read back the output in a way different from the generation path:
- Cross-verify whether the score for each dimension is supported by public sources
- Confirm that the 13 coverage rules have been applied correctly
- Check data freshness—whether data older than 90 days has been marked
- Verify that the rating is for the individual rather than the institution
- Confirm that No Rating cases have a clear exclusion reason
- Check whether the report contains unverified fabricated data
Step 6: Deliver
Output reports in the user's language (Chinese→Chinese, English→English). Clean up temporary files. Recommend that the user use investor-research to build a broader investor list after identifying target investors.
Output
Output reports using the following structure:
# [Investor Name] / [Firm] — Investor Due Diligence Report
Generated: YYYY-MM-DD | Perspective: [user's sector and stage]
## Rating: ⭐⭐⭐ / ⭐⭐ / ⭐ / No Rating (X.XX)
[One-line action recommendation]
| Key Metric | Data | Signal |
|-----------|------|--------|
| Dry Powder | Total AUM $X, est. remaining $X | 🟢 / 🟡 / 🔴 |
| Activity | X/yr, X in last 6mo | Very Active / Active / Conservative / Dormant |
| Sector Fit | X deals in your space in last 12mo | Precise / Broad / Adjacent / None |
| Decision Power | [Title], [X]yr at fund | Decision Maker / Influential / Advisory / Junior |
### Score Breakdown
| Dimension | Score | Basis |
|-----------|-------|-------|
| Dry Powder (25%) | X/4 | [one line] |
| Sector Fit (25%) | X/4 | [one line] |
| Decision Power (20%) | X/4 | [one line] |
| Activity (15%) | X/4 | [one line] |
| Lead Role (10%) | X/4 | [one line] |
| Risk (5%) | X/4 | [one line] |
| **Total** | **X.XX** | **⭐⭐⭐ / ⭐⭐ / ⭐** |
### Risk Flags
[List any red flags or write "None identified"]
### Bottom Line
[30-second decision summary]
Guardrails
Anti-patterns
- NEVER fabricate data—mark information that cannot be verified as "unverified" or downgrade its rating. Do not fabricate deal counts, AUM figures, or job titles
- Do NOT rate institutions instead of individuals—if only an institution name is given, ask the user which partner they met
- NEVER make personal attacks—focus only on investment behavior data, not gossip about personal life, divorce, or appearance
- Do NOT attempt to access private databases, paid APIs without a subscription, or non-public financial data
- NEVER deviate from the calibrated scoring weights—the 1-4 point scale and the 13 coverage rules are calibrated based on 120 real investor case studies
- Do NOT force a rating for every investor—"No Rating" is a valid output, with a clear reason for exclusion
Constraints
- Ratings target individuals, not institutions
- Language matching—always reply in the language the user uses, and do not mix Chinese and English unless quoting names
- Data older than 90 days must be marked with a freshness warning
- Founder time is the scarcest resource—every recommendation must answer: "Is this meeting worth my time?"
- Do not make personal attacks; focus only on investment behavior data
- Respect data boundaries; do not access private databases or non-public financial data
Related Skills
- investor-research — Research potential investors broadly to build a target investment institution list
- pitch-deck-creator — After identifying the investors to meet, create a professional pitch deck as meeting material
- deck-web-converter — Convert the pitch deck into a shareable web HTML page for easy sending before the meeting
About UniqueClub
This skill is part of the UniqueClub founder toolkit. Discover more at 🌐 https://uniqueclub.ai