Task `teams_ui_state_snapshot` v1

Readonly conceptual snapshot of the Teams global-search surface, distilled to a compact, schematic form (a single phase label + a few enums/booleans + short issues flags). No raw DOM/HTML and no free text from the page —…

Updated

Task teams_ui_state_snapshot v1

Summary

Readonly conceptual snapshot of the Teams global-search surface, distilled to a compact, schematic form (a single phase label + a few enums/booleans + short issues flags). No raw DOM/HTML and no free text from the page — designed as token-efficient facts for a governed decision break-in (automation_situation_arbitrate), including low-parameter models.

Inputs

  • session_ref (required)

Output

{ srf, phase, on_search, msg_tab, query, date, results, issues[] }

  • phase — one conceptual label for where the surface is: off_search, corrupted_query, wrong_tab, no_results, results_ready, search_ready_empty, unresolved, unknown.
  • query — schematic search-box state: wildcard | drifted | empty | text (never raw text).
  • dateset when a concrete (year-bearing) date chip is applied, else none.
  • resultssome when result cards are present, else none.
  • issues — short flags: off_search, query_drifted, wrong_tab, no_results_banner, no_date_scope, session_unresolved, state_unreadable.

Verification

tests/playwright/test_teams_ui_state_snapshot.py

Changelog

  • v1 — Compact schematic conceptual DOM snapshot for the v4 decision break-in.