pre-existing on the itd-183 branch: a compact nested mapping in a block sequence leaks an excluded key, and rawHTMLHeading's fence comment overclaims

pre-existing on the itd-183 branch: a compact nested mapping in a block sequence leaks an excluded key, and rawHTMLHeading's fence comment overclaims

Found by the round-9 adversarial security review of build/itd-183, and recorded rather than chased: identical on HEAD and on the parent 044ac6ed. The package does not exist on main, so nothing here is inherited from main.

1. A compact nested mapping in a block sequence leaks. In frontmatter,

   items:
     - origin: <value>

is a real origin key to YAML, but excludedKeyLineRe's ^\s* cannot cross the - , and flowKeyRe needs a { or a ,. The reviewer notes this shape is MORE LIKELY TO BE TYPED BY ACCIDENT than any of the round-9 regressions -- which makes it the most probable real-world leak on the branch, despite being the least exotic.

1. rawHTMLHeading's doc comment claims "a fenced line is replaced by an empty line rather than dropped". The code joins all lines unmodified and tests fenced[line] only for the OPENER, so a fenced region can still supply a bound for an unfenced heading (and, per iss-2608301237458660, a mask opener). The comment describes a behaviour the code does not have.

Both are pre-existing on the branch and are left open for the facilitator. Item 1 in particular deserves a decision: it is a genuine hole in the exclusion floor that no round has closed.

Grounds

• pursued: every compact nested mapping a frontmatter line can open behind block indicators refuses the assembly, while committed flow-mapping history rows and scalar sequences are admitted; an excluded key admitted in any such spelling, or a committed record refused by the new rule, would show it wrong