intake: read task dates and render the dashboard's Recent table (#249)

The Mind now dates every task the moment it leaves the backlog (PyAutoMind
REFERENCE.md "Task dates"). The census reads those dates back and the dashboard
grows the one section laid out by DATE rather than by state.

- parse_registry reads each entry's date via the key that names its event
  (`issued:` / `parked:` / `filed:`, most-specific first); a date sitting in
  another field's prose stays invisible, since that is the un-parseable habit
  the convention replaced. parse_header reads a prompt's own `Issued:` copy as
  the fallback that keeps an orphan in-flight prompt dated.
- completed_records walks complete/<YYYY>/<MM>/ newest-first and stops once it
  has enough — the ledger is 1000+ files and a 20-row table must not open all
  of them on every render. A record with no `completed:` field dates by the
  month folder the lifecycle engine filed it into.
- Recent sits between Backlog and Epics in both the markdown page and its
  Pages twin (a real table, with a copy button per row on the HTML side).
  Live work is SELECTED first and completions fill the rest: a straight date
  sort on a Mind that ships two hundred records a month is twenty receipts and
  no work, which would hide the dates on the very page they were added for.
- In flight / Parked / Planned rows now carry their date as a facet too.


Claude-Session: https://claude.ai/code/session_01L4HPWjv5rdzBAkKpfW1SbR

Co-authored-by: Claude <noreply@anthropic.com>

Author: Jammy2211

agents/conductors/intake/AGENTS.md

@@ -83,11 +83,22 @@ schema — light structure over free-form prose.
 | **classify** | `intake "<text>"` / `intake classify --file P` | classify one raw input; `--apply` writes the prompt |
 | **ideas** | `intake ideas` | scan `ideas.md`, propose one prompt per bullet; `--apply` writes them |
 | **census** | `intake census` | inventory every filed prompt (work-type/target/difficulty/status + hygiene flags); always read-only |
-| **dashboard** | `intake dashboard` | render the census as the Mind **task** page — picks, in flight, parked, planned, backlog; `--apply` writes `PyAutoMind/dashboard.md`, `--check` exits 1 on drift |
+| **dashboard** | `intake dashboard` | render the census as the Mind **task** page — picks, in flight, parked, planned, backlog, recent, epics; `--apply` writes `PyAutoMind/dashboard.md`, `--check` exits 1 on drift |
 | **formalise** | `intake formalise [prefix]` | retroactively header the prompts census flags — derive the missing fields, insert in place, prose untouched; `--apply` writes |
 | **reconcile** | `intake reconcile [prefix]` | rank backlog prompts that look already-shipped (vs the `complete/` records / `active/`); always read-only — retiring stays human |
 | **reconcile --repo** | `intake reconcile --repo <target> [prefix]` | **also** read the target repo's source for identifiers the prompts name — the one signal that sees a prompt with no Mind-side trace. Opt-in; the default path is offline |
 
+**Recent** is the one section laid out by *date* rather than by state: the 20
+newest task events merged across every bucket — issued, parked, filed,
+completed — sitting between the Backlog and the Epics. Every other section
+answers "what should I do now?"; recency is orthogonal to state, so none of
+them can answer "what has been happening?". Dates come from the registry key
+that names the event (`issued:` / `parked:` / `filed:`, PyAutoMind REFERENCE.md
+"Task dates"), a prompt's own `Issued:` header, or a record's `completed:`.
+Live work is selected first and completions fill the rest — on a Mind that
+ships two hundred records a month a straight date sort is twenty receipts and
+no work.
+
 Census/dashboard are the Mind *backlog* view — deliberately distinct from
 Heart's `/health status` health view (see "must never do"). The prompt-taxonomy
 folder is authoritative for a prompt's work-type/target; header fields are