Ingest coverage viz: git-commit history vs configured time-window (detect missing/at-risk activity)

Dispatch

canonical · plan

Spec

markdown

hand-off · dispatch

Auto-dispatch

when it reaches planned

Design-loop

design pass before build

Direct dispatch — no refine required. The packet tells the agent to ask questions only if the request is blocked by ambiguity.

target project worktree · optional branch · optional agent tool · optional session id · optional

kind

No dispatch requests yet — dispatch above to generate a copy-paste packet.

provenance · append-only

live

note added 1d ago

Spun out of #174 (operator Q&A #38, 2026-07-04 — "yes, separate brief"). Idea (from the operator's #174 note): Use historical git-commit ingest (flower:ingest-commits already exists) to chart a project's ACTUAL activity (commits over time) against its CONFIGURED ingest time-window — so we can SEE whether the window is missing, or would miss, real project activity. Makes "is my time-window too narrow?" answerable at a glance. Companion idea from the same note: ALSO pre-ingest SESSIONS (metadata only — do NOT summarize/embed/push-to-Meili until indexed) so sessions are visualizable relative to the configured window even before full ingest; this disambiguates "indexed but awaiting sessions" = genuinely no sessions in window vs window-too-narrow. Relation: complements #174 (ingest-health classification) and #112 (ingest visibility). Needs its own refinement (data sources: commits table + session started_at vs window config; the viz surface; whether the session metadata pre-ingest pass is in-scope or its own brief). NOT yet scoped/planned.

agent · flower-refine
participant joined 1d ago
system · flower-refine

epic · dependencies

epic parent

depends on

No dependencies — dispatchable once planned.

agents · waves

trace · graph

No links yet — they accrue as agents work the brief.

scope

dogfood · read-only

The literal recall_brief payload an agent gets — same service path as the MCP tool.