llm-security/tests/lib
Kjell Tore Guttormsen 90f576f056 fix(llm-security): v8.x-A - close <script> recall hole, add whole-table ReDoS gate
Two of the three confirmed v8.x-A evasions are closed; the third (attribute
padding) stays open by operator decision and is documented, not silently left.

Recall fix. hybrid-xss required a closing </script>, so `<script>alert(1)` and
`<script src=x.js>` both passed scanForInjection() with found: false, while the
closed form returned high. A src= tag has no body to close in the first place.
The opening tag alone is the signal, and matching it is strictly linear: one
negated-class run whose excluded character is its own terminator, so there is
no backtracking surface that could re-introduce v7.8.3 #24.

ReDoS gate. The #24 test covered six html-obfuscation patterns against the two
shapes that defect was found on. A catastrophically backtracking regex added
anywhere else in the four tables would have failed no test at all - which is
why #24 had to be found by hand. The new gate times every exported pattern and
asserts its own coverage, so it cannot be narrowed silently.

It has two classes because the blowup shapes need opposite inputs, and because
a synchronous RegExp.test() cannot be interrupted: an exponential pattern met
with a 512KB input would HANG the run rather than fail it. Class 1 uses a
28-char ambiguous-run ladder, where exponential costs ~1s and anything sane
costs microseconds; on a hit the 512KB sweeps refuse to run. Class 2 uses the
hook's real read cap to catch the polynomial #24 class. Proven to fire: an
injected exponential regex was named by label and the hang guard held.

Measured clean: worst single pattern 17ms, full large sweep 218ms.

Golden gate went red as expected and was re-blessed - diff is exactly two
lines, the regex source and the source-file digest. The 61-payload reference
run is byte-identical, so no showcase behaviour moved.

Suite 2073/2073 (was 2063; +5 recall tests, +5 gate tests).

Still open, unchanged: all 7 bounded HTML patterns evade on >256 chars of
padding, in two positions - before the attribute and inside the style value.
Wider than STATE recorded, which named only aria-label.
2026-08-10 14:53:29 +02:00
..
audit-trail.test.mjs feat(llm-security)!: v8 Phase 3 step 1 - remove the deprecated mode env-vars 2026-08-09 10:17:47 +02:00
bash-normalize.test.mjs feat(ultraplan-local): v1.6.0 — /ultraresearch-local deep research command 2026-04-08 08:58:35 +02:00
build-zip.mjs feat(llm-security): /security ide-scan <url> — Marketplace/OpenVSX/direct VSIX (v6.4.0) 2026-04-17 17:16:26 +02:00
commons-loader.test.mjs fix(llm-security): commons-loader - drop policy-driven root, ship path, fix cache [skip-docs] 2026-08-09 14:11:56 +02:00
diff-engine-exact-pass.test.mjs fix(llm-security): misc scanner correctness — reflog FP, diff exact-pass, hex dedupe, SARIF version (#20,#22,#50,#54,#56) 2026-07-18 10:36:51 +02:00
distribution-stats.test.mjs feat(ultraplan-local): v1.6.0 — /ultraresearch-local deep research command 2026-04-08 08:58:35 +02:00
doc-consistency.test.mjs docs(llm-security): trim README badge row 8 -> 4 (repo-standard BADGE-COUNT) 2026-08-04 11:50:12 +02:00
file-discovery.test.mjs fix(llm-security): normalization/discovery evasion + SIG embedded-base64 & custom rules (#21,#23,#30,#36,#42,#52,#55) 2026-07-18 10:35:56 +02:00
git-clone-gitattributes.test.mjs feat(git-clone): E12 — .gitattributes filter-driver post-clone advisory 2026-04-30 15:29:13 +02:00
git-clone-sandbox.test.mjs feat(ultraplan-local): v1.6.0 — /ultraresearch-local deep research command 2026-04-08 08:58:35 +02:00
golden-baseline.test.mjs fix(llm-security): golden gate - coverage block measured something else 2026-08-09 13:06:45 +02:00
injection-patterns.test.mjs fix(llm-security): v8.x-A - close <script> recall hole, add whole-table ReDoS gate 2026-08-10 14:53:29 +02:00
mcp-description-cache.test.mjs fix(llm-security): scanner robustness — ReDoS, MCP-stdout DoS, redirect loop, atomic writes (#24,#53,#31,#25,#51) 2026-07-18 10:15:11 +02:00
output.test.mjs feat(llm-security)!: v7.0.0 commit 6 — tests, docs, version bump 2026-04-19 22:26:35 +02:00
policy-loader.test.mjs feat(llm-security)!: v8 Phase 3 step 1 - remove the deprecated mode env-vars 2026-08-09 10:17:47 +02:00
severity.test.mjs docs(severity): B3 — document info as scoring-inert (v7.2.0 prep) 2026-04-29 13:56:11 +02:00
skill-registry-atomic.test.mjs fix(llm-security): scanner robustness — ReDoS, MCP-stdout DoS, redirect loop, atomic writes (#24,#53,#31,#25,#51) 2026-07-18 10:15:11 +02:00
string-utils-hidden-unicode.test.mjs feat(unicode): E1 — extend hidden-Unicode detection to PUA-A and PUA-B 2026-04-29 14:18:49 +02:00
string-utils-homoglyph.test.mjs feat(injection): E16 — homoglyph NFKC fold before every pattern match 2026-04-29 14:22:05 +02:00
string-utils-tokens.test.mjs fix(dep): B7 — token-overlap typosquat heuristic alongside Levenshtein 2026-04-29 14:10:53 +02:00
string-utils.test.mjs fix(llm-security): normalization/discovery evasion + SIG embedded-base64 & custom rules (#21,#23,#30,#36,#42,#52,#55) 2026-07-18 10:35:56 +02:00
v8-env-removal.test.mjs feat(llm-security)!: v8 Phase 3 complete - riskScoreV1, posture heuristic, docs 2026-08-09 10:25:03 +02:00
workflow-yaml-state.test.mjs fix(llm-security): YAML/workflow parser divergence — block scalars + bare if: (#32,#33,#43) 2026-07-18 10:35:56 +02:00
yaml-frontmatter.test.mjs fix(llm-security): YAML/workflow parser divergence — block scalars + bare if: (#32,#33,#43) 2026-07-18 10:35:56 +02:00