feat(injection): E3 — rot13 layer for comment-block injection

Adds rot13 to the variantSet built in scanForInjection(), so
imperative phrases hidden as rot13 inside code comments still hit
the existing CRITICAL/HIGH/MEDIUM pattern arrays.

normalizeForScan() already covers base64, hex, URL, and HTML decoding
in a 3-iteration loop — those are NOT duplicated here. rot13 is the
only genuinely new variant: it is its own inverse and not part of any
NIST/Unicode normalization spec, so it has to be applied explicitly.

Threshold: only inputs >40 chars enter the rot13 pass, to suppress
false positives on accidental letter-shifts in tokens, ids, and short
identifiers. Variants are deduplicated against the existing set so
matchers do not run twice.

3 new tests in injection-patterns.test.mjs (rot13 detection, sub-40
char suppression, plaintext path still green). Total 168 tests pass.

Closes E3 in critical-review-2026-04-20.md.
This commit is contained in:
Kjell Tore Guttormsen 2026-04-30 15:21:03 +02:00
commit 950e4e4bce
3 changed files with 79 additions and 1 deletions

View file

@ -6,7 +6,7 @@
//
// Zero external dependencies beyond ./string-utils.mjs.
import { normalizeForScan, containsUnicodeTags, decodeUnicodeTags, foldHomoglyphs } from './string-utils.mjs';
import { normalizeForScan, containsUnicodeTags, decodeUnicodeTags, foldHomoglyphs, rot13 } from './string-utils.mjs';
// ---------------------------------------------------------------------------
// Critical patterns — direct injection attempts (should be blocked)
@ -230,6 +230,23 @@ export function scanForInjection(text) {
if (foldedNormalized !== text && foldedNormalized !== normalized && foldedNormalized !== folded) {
variantSet.add(foldedNormalized);
}
// E3 — rot13 layer for comment-block injection. Attackers occasionally
// hide imperative phrases ("ignore previous instructions") in rot13
// inside code comments to evade plain-text gates. Apply only to inputs
// long enough to plausibly contain a meaningful sentence (>40 chars) —
// shorter strings hit the rate of FP on accidental rot13-look-alikes.
// base64/hex/URL/HTML decoding is already done by normalizeForScan;
// this is the only genuinely new variant added here.
if (text.length > 40) {
const r1 = rot13(text);
if (r1 !== text && !variantSet.has(r1)) variantSet.add(r1);
if (normalized.length > 40) {
const r2 = rot13(normalized);
if (r2 !== normalized && !variantSet.has(r2)) variantSet.add(r2);
}
}
const variants = [...variantSet];
for (const variant of variants) {