open/llm-security
1
0
Fork 0
Code Issues Releases 1 Activity

fix(llm-security): close OpenAI legacy key recall gap in pre-edit-secrets hook

Browse source 
Bare/unquoted legacy OpenAI keys (no label assignment, no Bearer prefix)
slipped past the pre-write secret-detection hook. Added a pattern anchored
on the T3BlbkFJ base64 "OpenAI" watermark (vendor-documented shape),
avoiding the collision-prone bare sk-+48alnum form. Failing tests first,
full suite green (2184/0/6).

The originally planned source for this fix — porting two entries from
commons' secret-egress.json — turned out to be a false premise: that file
is a byte-identical copy of this hook's own table, not a superset. The two
missing names existed only as prose in commons' conformance/manifest.json,
describing a different repo's (the guard's) unpublished Python table.
gcp-service-account-json was measured NOT to be a gap (already covered by
the existing PEM-block pattern); openai-api-key-legacy was the one real
gap, closed here with a locally-authored pattern rather than an invented
"port". Commons notified via coord-send that their secret-egress.json
(count: 18) is now stale.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PYJX35KLH3rpS6pi7LHj8u
This commit is contained in:
Kjell Tore Guttormsen 2026-08-11 13:43:27 +02:00
commit 088e45836c
2 changed files with 50 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

44
tests/hooks/pre-edit-secrets.test.mjs
View file

@ -59,6 +59,14 @@ const jwtToken = [
'.', 'TJVA95OrM7E2cBab30RMHrHDcEfxjoYZgeFONFh7HgQ',
].join('');
// OpenAI legacy API key: sk- + 20 chars + the base64 "OpenAI" watermark
// (T3BlbkFJ) + 20 more chars. The watermark anchor is what the shape check
// keys on, per vendor-documented format (avoids the bare sk-+48alnum shape,
// which collides with sk-ant-/sk-proj- and other unrelated sk-* tokens).
const openaiLegacyKey = [
'sk-', 'a'.repeat(20), 'T3BlbkFJ', 'b'.repeat(20),
].join('');
// ---------------------------------------------------------------------------
// Helpers
// ---------------------------------------------------------------------------
@ -216,6 +224,42 @@ describe('pre-edit-secrets — bare provider keys (#13)', () => {
});
});
// ---------------------------------------------------------------------------
// OpenAI legacy API key — recall gap. A bare/unquoted legacy key (no label
// assignment, no Bearer prefix) previously slipped through: sk-ant-api03-
// and sk-proj- are covered, but the pre-2024 sk-<48 chars> shape had no
// dedicated pattern and only 'Generic credential assignment' or
// 'Authorization header with token' happened to catch it in labeled/header
// contexts — not when it appears bare, e.g. an unquoted env assignment.
// ---------------------------------------------------------------------------
describe('pre-edit-secrets — OpenAI legacy key recall gap', () => {
it('blocks a bare legacy OpenAI key in an unquoted env assignment', async () => {
const result = await runHook(SCRIPT, writePayload(
'.env',
`OPENAI_API_KEY=${openaiLegacyKey}`
));
assert.equal(result.code, 2);
assert.match(result.stderr, /BLOCKED/);
assert.match(result.stderr, /OpenAI/);
});
it('blocks a bare legacy OpenAI key with no surrounding context at all', async () => {
const result = await runHook(SCRIPT, writePayload('notes.txt', openaiLegacyKey));
assert.equal(result.code, 2);
assert.match(result.stderr, /BLOCKED/);
assert.match(result.stderr, /OpenAI/);
});
it('allows prose mentioning the sk- prefix without the T3BlbkFJ watermark', async () => {
const result = await runHook(SCRIPT, writePayload(
'docs/notes.md',
`Legacy OpenAI keys start with sk- followed by 48 characters, e.g. sk-${'x'.repeat(48)}.`
));
assert.equal(result.code, 0);
});
});
// ---------------------------------------------------------------------------
// ALLOW cases
// ---------------------------------------------------------------------------