# SKSecurity honest-claims allowlist
# ----------------------------------------------------------------------
# One fnmatch glob (or path substring) per line. Files matching any entry
# are SKIPPED by the no-overclaim gate. Comments start with '#'.
#
# Use this ONLY for files that must contain raw forbidden phrases as data
# (test fixtures, the scanner's own pattern table) — NOT to hide real
# marketing overclaims. Honest-discipline docs (SOP, SECURITY, etc.) do
# NOT need entries here: they quote/negate the forbidden tokens, which the
# scanner already recognises as meta-references, not claims.

# Test fixtures: deliberately assert raw overclaims to prove detection works.
tests/test_honest_claims.py

# Audit reports: reproduce the flagged phrases verbatim as evidence (the
# findings they report), same data-file rationale as the test fixtures above.
*audit-*.md
