v1.32.0
[1.32.0] - 2026-09-06
Every room's moves become one report. A new script reads the transactions pulled from
each of the user's leagues and prints what every manager added, dropped and swapped, each
player dropped in the window with an estimated clear date (his drop date plus that
league's waiver period), and the names the market chased in more than one league. Section
8 says what the copilot does with each of those, on waiver eve and on demand.
Added
scripts/transactions.py(Python 3.8 or later, nothing else). It reads a private
folder holdingleagues.mdand one pipe-row file per league, de-duplicates the rows
that repeat when pulls overlap, and prints four sections to standard output: the
per-league manager digest (adds and drops by position, swaps, last move date), every
player dropped in the window with what he was swapped for in each league that dropped
him and an estimated clear date there (the drop date plus that league's waiver period,
arithmetic and not the host's own processing), the players added in two or more
leagues, and the user's own adds and drops. A dropped player the same league has since
added says so in his entry.--diris required,--sincedefaults to seven days
before the newest row the run read, and--leaguelimits the run to one slug. Any line
the shape cannot read stops the run with the file and the line number, the txn cell
included: it is the key that pairs a drop with the add that made room for it, so it is
required, is never-, belongs to one timestamp, and carries one manager's adds and
drops, a trade's two rows excepted. The script surfaces what the rooms did; it does not
rate, rank or recommend.references/transaction-watch.md: what a host's transactions page shows, the browser
pull method, the shape of the two files, exactly what each section of the report holds,
and the command. It is a new file rather than a section of
references/field-behaviors.md, which holds the live draft-day observations behind
section 5 and not a pull method or a file format.
Changed
- Section 8 gains one bullet, "Transaction watch", after "Waiver eve": the pull and the
run, then what each part of the report is for. Every dropped name is a candidate and
takes the full read plus this section's replaceability and burn-the-claim tests, whose
reading of the host's own row governs the real cost and clear date, the report's date
being arithmetic on the drop date; a dropped player already picked up in that league is
shown as such; a drop
paired with a same-position add in one transaction is a role or injury question for the
knowledgebase before anything else; a dropped streaming defense or kicker is noise
unless the pool is thin; a name added in two or more leagues is the market's read, a
data point and never a directive; and the manager digest reads each room's needs,
feeding the bye-week stretch's trade targets and the burn-the-claim test's contention
input. - README counts eleven shipped scripts and names the new one.
Why: the drop side of the wire is where the other rooms' mistakes land, and the copilot
was reading one league's free agent list rather than every league's transaction page. No
rule about how a candidate is judged changed; the report only puts the names in front of
the tests that already existed.