CVE-2022-3602:OpenSSL: Off-by-one stack buffer overflow in X.509 name-constraint pu
A reconstruction of the flaw, and the evidence path Lachesis follows to make it visible.
With no family and no line seeded, the whole-taxonomy enumerator put 3 of its memory.copy.capacity leads inside crypto/punycode.c (lines 187/276/297) -- the.
That is the shape of the failure. The rest of this case file follows the evidence behind it.
What was known.
What Lachesis established.
Known before Lachesis ran
The CVE and vulnerable release were selected from public history. We are not claiming novel discovery.
- CVE
- CVE-2022-3602
- Vulnerable target
- openssl-3.0.0 (last-vulnerable; range 3.0.0-3.0.6) · vulnerable
- Prior knowledge
- CVE-2022-3602 · openssl-3.0.0 (last-vulnerable; range 3.0.0-3.0.6) is vulnerable
Independent re-detection
The experiment graph was built around the last-vulnerable release (openssl-3.0.0), scoped to crypto/punycode.c plus include/crypto and include/internal so the single vulnerable translation unit builds under the pass budget (build 1s, enrich 1811 nodes / 1952 edges), then enumerated BLIND over the whole constructor registry (no family hard-coded, no target line seeded). The census and candidate ids below are the enumerator's own verbatim output; the graph was released immediately after capture, so the provenance body-reads are not re-quoted here. Prose in `adjudication`/`trace` is MY reading of the public advisory + fix, clearly separated from tool output.
- Seeded inputs
- None
- Run timestamp
- 2026-09-03T00:00:00Z
- Evidence artifact
- cve-2022-3602.kuzu (built from OpenSSL openssl-3.0.0, scoped to the punycode translation unit; graph + sidecars deleted after capture)
The commands and outputs below come from this recorded Lachesis run.
What Lachesis reconstructed.
One historical repository. Four captured queries. A complete source-to-sink argument.
Lachesis follows the same evidence path through unfamiliar repositories.
Here is how the path becomes visible.
These are the recorded questions Lachesis asked of the historical vulnerable code, followed by the raw result and source location each query returned.
Captured runA real replay of Lachesis over the graph we built for this case (build 1s under the pass cap, bounded memory). Every value below is verbatim captured tool output; the graph was released immediately after capture, so nothing accumulates.
Investigate the candidate
The narrowed scope builds the single vulnerable TU in 1s under the pass cap (the prior all-of-include/ scope tripped a manifest/snapshot node mismatch; the two-header-dir scope is clean).
build rc=0 1s
enrich rc=0 :: pass2 published 1811 nodes and 1952 edgesA small line with a large consequence.
the index-vs-capacity comparison guarding the label write is off by one, so exactly one element is written past the fixed buffer for a crafted label length.
// decoded label element written to the output buffer with an off-by-one capacity bound
// 3.0.7 corrects the bound so the output index is checked against the true remaining capacity before every write, and fixes ossl_a2ulabel's size accounting.
The original record.
Read the historical advisory and vulnerability record behind this reconstruction.