[mabry1985]

Summary

The Ready gate's only pre-failure nudge toward smaller slices was the breadth cap, which counts files_to_modify — so a conceptually wide card with few files (ds-syt / protoContent#609: 4 files, medium, timed out twice at 1800s) slipped through. This adds a non-blocking spec-width advisory in _ready_findings that reads width out of the spec itself via three tunable signals — spec length over SPEC_WIDTH_MAX_CHARS (2500), acceptance-criteria count over SPEC_WIDTH_MAX_CRITERIA (6, counted with the existing _decompose_ac so it matches the requirement ledger), and "and also"-style concern joins over SPEC_WIDTH_MAX_JOINS (3). The message names every signal that tripped with its measured value and threshold, and the fix suggests splitting into one-concern cards.

It only ever appends to advisories (never refusals), so ready_check(...)["ok"] stays True and mark_ready still promotes a wide-but-valid card. It skips task-type beads and runs before the early return, so a card with no files and no depends_on is still measured. Because create/update already return ready_check, the advisory surfaces at board_create_feature / board_update_feature automatically. New constants live beside MAX_FILES_BY_DIFFICULTY; docs/lifecycle.md documents the advisory and its thresholds; changelog.d/548.md records the change.

Fixes #548

[protoreview]

QA panel review — PASS

code-review-structural · head 2f1cc8e859c5 · formal

No findings were raised, so there was nothing to verify. The change is purely additive (+269/-0), introducing a spec-width advisory to _ready_findings. All five review angles found no defects. The protoPatch engine was unavailable in this environment; the LLM structural finder covered that lane manually and found nothing. No action required.

No findings — the review came back clean.

findings JSON (machine-readable)

[]

[protoreview]

Promoting the PASS verdict for head 2f1cc8e859c5: all checks terminal-green, zero unresolved review threads. (approve-on-green)