score-vc-fit
Skill4 starsChanged 8 months agoMITView on GitHub
Score and rank VC firms and target investors based on fit with the startup's fundraising brief. Uses a weighted 8-dimension rubric. Produces a prioritized outreach list with clear reasoning. Triggers on: "score vcs", "rank investors", "evaluate fit", "step 5", "next step", "continue to score", "prioritize investors".
In this repository8 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- clarify-thesisSkill
- draft-investor-outreachSkill
- enrich-vc-contactsSkill
- export-csvSkill
- find-vcsSkill
- map-investorsSkill
- resetSkill
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 4 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 22
Read all three files:
- Distinctive: line 27
If any file is missing, stop and tell the founder which step to run first.
- Very specific: line 48
If evidence for a dimension is unavailable, score it null — do not score it 0
- Distinctive: line 51
If more than 3 dimensions are null, flag the record as low-confidence
- Distinctive: line 61
Do not reward prestige; reward actual fit evidence
- Very specific: line 67
Write an array to data/normalized/scored_funds.json conforming to schemas/fund_match.schema.json.
- Very specific: line 73
After writing data/normalized/scored_funds.json, present the ranked list to the founder:
- Very specific: line 101
Do not proceed to draft-investor-outreach until the founder confirms or adjusts the ranking.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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
Score VC Fit
Pipeline: clarify-thesis → find-vcs → map-investors → enrich-vc-contacts → score-vc-fit → draft-investor-outreach
You are step 5 of 6. Your job is to turn the candidate list into a prioritized outreach list using a structured scoring rubric.
Tools
None. This step is reasoning only — scoring based on data already collected in previous steps.
Input
Read all three files:
data/normalized/company_profile.json— the startup's fundraising profiledata/raw/fund_candidates.json— candidate firms with rationaledata/normalized/investor_contacts.json— enriched contacts
If any file is missing, stop and tell the founder which step to run first.
Scoring rubric
Score each fund out of 100 using 8 weighted dimensions:
| Dimension | Max points | What to assess |
| Sector fit | 25 | Does the fund actively invest in this exact sector/subsector? |
| Stage fit | 20 | Does the fund's sweet spot match the raise stage? |
| Geography fit | 10 | Is the fund active in the startup's HQ geography or customer geography? |
| Check size fit | 10 | Does the fund's typical check size match the target? |
| Lead behavior fit | 10 | Does the fund's lead/follow style match the founder's preference? |
| Portfolio adjacency | 10 | Has the fund invested in comparable companies or adjacent spaces? |
| Recency | 10 | Has the fund made investments in the last 24 months (not just deploying old fund)? |
| Partner relevance | 5 | Is the identified contact person well-matched to this deal? |
Score total = sum of non-null sub-scores, rescaled to 100:
score_total = (sum_of_non_null / score_max_possible) * 100
Missing evidence handling:
- If evidence for a dimension is unavailable, score it
null— do not score it 0 score_max_possible= sum of max points for all non-null dimensions- Note missing dimensions in
score_notes - If more than 3 dimensions are null, flag the record as low-confidence
Score interpretation:
- 90–100: excellent fit — prioritize first
- 75–89: strong fit — high priority
- 60–74: possible fit — worth pursuing if list is short
- below 60: weak fit — deprioritize or remove
Scoring rules:
- Be conservative — do not over-score weak evidence
- Do not reward prestige; reward actual fit evidence
- Penalize clear mismatches even if other dimensions are strong
- Distinguish "famous fund" from "right fund"
Output
Write an array to data/normalized/scored_funds.json conforming to schemas/fund_match.schema.json.
Sort by score_total descending. Join investor_contact objects into each fund record using fund_id.
Review checkpoint
After writing data/normalized/scored_funds.json, present the ranked list to the founder:
Ranked [N] VC firms for [company_name]:
| Rank | Fund | Score | Top contact | Sector | Stage | Concerns |
|------|------|-------|------------|--------|-------|---------|
| 1 | ... | 87/100 | ... | ✓ | ✓ | none |
| 2 | ... | 74/100 | ... | ✓ | ~ | check size unclear |
...
Score breakdown for top 5:
[Fund name] — [score]/[max_possible]: sector [N], stage [N], geography [N], ...
Score notes: [score_notes]
[If any funds scored below 60]:
⚠ [N] firms scored below 60 and will be deprioritized. Remove them entirely? (y/n)
[If many null dimensions]:
Note: [N] firms had incomplete evidence — scores may shift with more research.
Does this ranking look right? You can:
- Confirm to proceed to step 6 (drafting outreach messages)
- Move a firm up/down: "prioritize [fund]" or "deprioritize [fund]"
- Remove a firm: "remove [fund]"
- Ask for more detail on a score: "explain score for [fund]"Do not proceed to draft-investor-outreach until the founder confirms or adjusts the ranking.
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.

