0103 pastes a command whose output has moved from three files to five

[iderex]

0103 pastes a command and its output to show that three refusals in it have a
worked case behind them. The command returns five files today and the paste says
three, and the two it does not name are the record itself and the record that
wrote the case down first.

The evidence

git rev-parse origin/main
ded6d1fcbd9ad65b5a04fbf58c0a7d0e34dd2c10
git show origin/main:docs/decisions/0103-what-admits-a-dependency-and-what-is-refused.md | grep -A4 "git grep -l '#103'"
    git grep -l '#103' -- docs/decisions
    docs/decisions/0041-how-a-cache-key-is-built.md
    docs/decisions/0061-the-span-facility.md
    docs/decisions/0112-where-the-platform-decoder-begins.md

Re-run against the reference a reader has:

git grep -l '#103' origin/main -- docs/decisions
origin/main:docs/decisions/0011-the-language-the-toolchain-and-the-binding-layer.md
origin/main:docs/decisions/0041-how-a-cache-key-is-built.md
origin/main:docs/decisions/0061-the-span-facility.md
origin/main:docs/decisions/0103-what-admits-a-dependency-and-what-is-refused.md
origin/main:docs/decisions/0112-where-the-platform-decoder-begins.md

Five rather than three. One of the two the paste does not carry is 0103 itself,
which matches on its own Issue: line and always did, so that half of the
disagreement was there on the day the paste was written. The other is 0011, which
is the record that wrote this collision down as its own reversal condition before
there was a graph to measure, and it is the most useful of the five to a reader
following that sentence.

Why it matters here rather than as a tidy-up

The paragraph the paste sits under is arguing that a refusal with a worked case
behind it is quotable. A reader who runs the command finds two files the record
does not account for and cannot tell whether the record is wrong or whether they
are. That is the defect class this repository's own rules lead with, in a record
whose subject is what may be admitted at all.

It is also not repairable by editing the record. 0001 permits a typographical
fix, a superseded by line and a pointer, and a stale reading is none of the
three, so what closes this is a decision about how a record carries a reading
that moves - not a quiet edit to the block above.

Done when

The record no longer asserts a count that the command beside it contradicts, by
whichever route 0001 permits: a superseding record, or the paste replaced by the
command alone in the way docs/gate-parity.md and README.md already hand a
reader a command rather than its output. Naming the route is part of this issue
rather than assumed by it.

[iderex]

What repairs this is decided by #267 first, and adding blocked-on-issue for
that reason.

Every route out of here edits or replaces a landed record. 0001 permits three
edits and a stale reading is none of them, so the repair is either a supersession
or whatever shape #267 settles for a record that is wrong in one place and right
everywhere else. Superseding 0103 whole to correct one pasted grep would discard
a licence set, a worth test and five refused behaviours that nobody has argued
with, which is the same disproportion #267 exists to answer.

gh issue view 267 --repo Flowfin/core --json number,state,title --jq '"#\(.number) \(.state) \(.title)"'
267 OPEN A record that narrows one clause of another has no shape in 0001, and one now exists in the tree

So this is not a second decision. It is the first instance of #267's answer being
applied, and it is worth keeping separate because #267 could be answered in a way
that does not reach a stale reading at all.