fix(browser): narrow credential-query denylist to unambiguous names

Follow-up on the salvaged #49830 hardening. The contributor's sensitive
query-param set included bare English words (code, key, auth, session,
sig) that double as ordinary page facets — ?code= on promo/challenge
pages, ?key= as a search facet, ?session= on blogs — so web_extract and
cloud browser_navigate would refuse a large slice of normal browsing.

Narrow the set to unambiguously credential-named params (access_token,
authorization, client_secret, password, token, x-amz-signature, ...).
Prefix-based vendor-key redaction (is_safe_url) still catches recognizable
key shapes; this set is the belt-and-suspenders for opaque secrets carried
under an explicit credential-named parameter.

Also fixes two intra-PR-staleness test breakages surfaced by salvaging onto
current main:
- web_extract_tool() no longer accepts use_llm_processing= (signature
  changed since the PR was authored) — dropped the invalid kwarg.
- agent.redact now fully masks keyed 'token=<secret>' to 'token=***'
  instead of partial 'sk-...'; the console-redaction test now asserts the
  real invariant (secret body gone) rather than the exact mask format.

Added a regression test that generic English-word query params are NOT
blocked by the credential guard.
This commit is contained in:
teknium1 2026-07-01 04:47:58 -07:00 committed by Teknium
parent 937e56be92
commit cfbc7ed1f9
4 changed files with 39 additions and 10 deletions

View file

@ -112,9 +112,13 @@ class TestBrowserConsole:
result = json.loads(browser_console(task_id="test"))
serialized = json.dumps(result)
# The secret body must be gone. The exact mask format
# (partial ``sk-…7890`` vs full ``***`` for keyed ``token=`` values)
# is owned by agent.redact and intentionally not pinned here.
assert "BROWSERCONSOLESECRET" not in serialized
assert "sk-" in result["console_messages"][0]["text"]
assert "..." in result["console_messages"][0]["text"]
redacted_text = result["console_messages"][0]["text"]
assert fake_key not in redacted_text
assert "***" in redacted_text or "..." in redacted_text
def test_redacts_secrets_from_eval_result(self):
from tools.browser_tool import _browser_eval