1
0
Fork 0
llm-ingestion-pipeline-secu.../docs
Kjell Tore Guttormsen 45c2b06315 docs: "warn, clean" is graded, never cleaned bytes -- measured at okf's door
okf measured that 0.7.0's no-URL narrowing dispositions four raw-HTML
carrier forms (<a aria-label>, bare </a>, <Frame>, <video />) to warn
with no active-content finding -- verified here directly against
screen_output(..., PRESET_USER_UPLOAD), which returns the library's own
"clean: no findings" reason string (disposition.py:264). At their door
(llm-ingestion-okf 0.7.0), warn is the persist floor (inbox.py:139/323),
and their adapter forwards the original extracted text because
screen_output never hands back sanitized bytes -- defanging is a
separate, deliberate neutralize() call. So the same four forms land
written into the bundle, carrier present verbatim. Named per operator
decision; CHANGELOG.md's "35 limitations" and GATE-G-v1.md's counts are
dated snapshots of past states and are left as measured, not bumped.

35 -> 36: README.md's summary line is the only other live count.
2026-08-13 23:24:38 +02:00
..
ADOPTION-BRIEF.md release: 1.1.0 -- the first behaviour change shipped under the freeze 2026-08-13 22:58:36 +02:00
BRIEF.md release: 1.1.0 -- the first behaviour change shipped under the freeze 2026-08-13 22:58:36 +02:00
fp-sweep.py measure(fp): the upload door costs a human on 16-53% of benign documents 2026-08-10 21:32:56 +02:00
GATE-G-v1.md docs(gate): the surface claim covers okf too, and one branch of D4 is already a defect 2026-08-13 22:07:18 +02:00
LIMITATIONS.md docs: "warn, clean" is graded, never cleaned bytes -- measured at okf's door 2026-08-13 23:24:38 +02:00
OKF-INGESTION-BRIEF.md docs: version-sync + SECURITY/CONTRIBUTING + honest-limits (Session E) 2026-07-15 10:08:24 +02:00
PLAN-v1.md docs(plan): Session G landed, D4's rationale lives where the promise does, and the ninth surface was verified not bumped 2026-08-13 22:45:11 +02:00
PLAN.md fix(output): 19 quadratic regex runs on the output path, worst ~5.7h at the cap 2026-07-31 18:31:58 +02:00
rawhtml-census.py feat(active-content): raw HTML graded on carrier, and a tag naming no target is inert 2026-08-12 00:42:44 +02:00
redos-sweep.py fix(sanitize,okf,active_content): three quadratic patterns, two on the input path 2026-08-01 20:06:36 +02:00
review-2026-07.md docs: reframe novelty claim to composite write-time contract (review MAJOR #3) 2026-07-15 09:22:38 +02:00
URL-SHAPE.md docs(url-shape): make the rule reconstructable, and record what three corpora measured 2026-07-27 08:56:24 +02:00