The fastest way to find out whether an appraisal flatters its client is to point it at the appraiser. RepoAppraisal was run against Scottcjn, the GitHub account of its own author, on 4 August 2026. The production metrics came back top-of-market. The market-absorption metrics came back in the bottom quartile. Both are published below, and the unflattering half is first on purpose.
These are the findings that reduce the appraised value. An appraisal that leads with strengths and buries the defects in an appendix is a brochure, not an appraisal.
Across 150 non-fork repositories the account holds 6,662 stars — about 61 stars per repository. That places it 18th out of 19 in its comparable set on stars-per-repo, while placing 1st out of 19 on annual contribution volume. The account ships more than anyone it was compared against and converts less of it into attention than almost anyone it was compared against.
Verdict: over-improved, under-marketed.The Rustchain repository shows an approximately 1:1 star-to-fork
ratio. The companion rustchain-bounties repository has more forks
than stars. Ordinary open-source repositories run somewhere between 10:1 and 50:1 stars
to forks. A ratio at or below parity is the signature of a paid contribution funnel: people are
arriving to do bounty work, not to bookmark a project they admire.
This is disclosed and adjusted under the terms-of-sale adjustment, and it is reported as a separate class of attention. It is not called fraud, and no accusation of fake accounts or purchased stars is made or implied. A house that sells with seller concessions is still a real sale; the appraiser discloses the consideration and adjusts for terms. Incentivized attention is a real transaction with a known price attached.
Verdict: real conversion, disclosed consideration, adjusted — not blended into organic demand.GitHub reported 12,840 clones against only 3,891 page views in the same 14-day window — a 3.3:1 inversion. Human beings look before they clone, so healthy repositories run the other way round. Sixty-six repositories recorded zero page views but non-zero clones, which is the fingerprint of CI runners, mirrors, package indexes, and scraping bots rather than developers.
The clone number was therefore discarded as an adoption proxy. The honest adoption figure used in this appraisal is 1,890 unique visitors over 14 days.
Verdict: clone counts rejected as evidence of adoption.67 of 150 repositories — 45% — received zero human page views in the 14-day measurement window. The top 5 repositories carry 78% of all views. The portfolio is not 150 assets; it is roughly five assets and a long tail of published work that nobody is currently reading.
Verdict: extreme concentration; the tail carries option value at best, not income.92.6% of referred traffic originates from github.com itself plus Google search. There is no meaningful independent referral base — no newsletter, no forum, no conference circuit, no publication funnel sending readers in. The attention the portfolio receives belongs to the platform and the search engine, not to the author. It does not travel if the author leaves the platform.
Verdict: audience is platform-dependent and non-transferable.471 distinct external contributors appear in the record. Of those, 91.8% appear in exactly one calendar month and never return. Zero contributors lasted four months. Only three lasted three months or more.
The fork evidence agrees. 0 of 106 forks carry their own commits on their default branch, and 52% of fork branches are named after a bounty number. That is a labour market clearing transactions, not a community forming around a project.
Verdict: a working bounty machine, not a community. Priced as throughput, not as goodwill.The adverse findings above are real. So is everything in this section. An appraisal reports both or it reports nothing.
72 pull requests merged into repositories the account does not control. This is the metric that cannot be farmed: it requires a stranger with commit rights to say yes.
Trailing twelve months. Ranked first in the comparable set, ahead of every account it was measured against.
542 followers places the account in the top 0.014% of 155,571,951 GitHub accounts. 222 public repositories is the top 0.039%.
Two commits authored by this account are present in openssl/openssl:
e443447 and f56a9fa. OpenSSL's maintainers cherry-pick contributions and
then close the originating pull request, so the pull request status field reads
closed rather than merged. The commits, not the pull request status, are
the record. Any appraisal that scores merge credit purely off the GitHub
merged boolean would score this contribution at zero, which is wrong. A merged pull
request in amd/xdna-driver (#1448) is also confirmed.
Published packages on PyPI and npm draw approximately 5,000 combined downloads per month. Download counts include mirrors and CI, so this figure is treated as an upper bound on human usage, not a user count.
Nineteen accounts in adjacent systems-programming and LLM-infrastructure niches. Only public, factual, machine-readable metrics are shown. No valuation, ranking of worth, or authenticity judgment is attached to any named third party.
| Account | Contributions (12 mo) | Rank in comp set |
|---|---|---|
| Scottcjn (subject) | 6,044 | 1 of 19 |
| ggerganov | 4,119 | 2 of 19 |
| geohot | 2,929 | 3 of 19 |
Three approaches were run and reconciled. The result is a range, not a number, because the evidence supports a range and not a number.
This is an illustrative valuation framework, not a certified appraisal. It is not prepared under USPAP, Elyan Labs LLC is not a licensed appraisal firm, and this figure must not be used for lending, tax, employment, litigation, or any transaction where a certified valuation is required or expected.
The 2:1 spread between the low and high end is itself a finding. The cost approach — what it would cost to reproduce the tracked output and the physical plant behind it — supports the upper end. The income approach, capitalized against actual attributable revenue, supports the lower end. When production evidence and absorption evidence disagree this sharply, an honest reconciliation widens the band instead of splitting the difference and calling it precision.
Read as an appraiser reads a property file: this is a well-built structure on a street with no foot traffic. The improvements are real and expensive to reproduce. The absorption is poor. Those are separate facts and they stay separate in the report.
The single most actionable finding is #4. Forty-five percent of the portfolio is generating no measurable human interest. Consolidating, archiving, or deliberately marketing that tail would move the appraised range more than shipping another repository would. The account's constraint is not production capacity; it is distribution.
The second most actionable finding is #6. A bounty funnel that converts 471 people and retains none of them is a functioning labour market and a failed community. Those are priced differently. Throughput can be bought again next month. Goodwill compounds. Only one of them was built here.