feat: intake census + dashboard modes — Mind backlog page

[Jammy2211]

Overview

Add the two follow-up modes the Intake Agent's own docs already name as planned: intake census (a read-only inventory of every filed PyAutoMind prompt) and intake dashboard (renders that census as a clean backlog page committed to the PyAutoMind repo). This gives the Mind a single GitHub-rendered view of the backlog — what is filed, where, at what difficulty/priority — plus hygiene flags for malformed prompts. The repair sibling mode stays out of scope.

Plan

• Add intake census: walk all work-type folders + triage/ in PyAutoMind, parse each prompt's light header, and report an inventory with hygiene flags (missing header fields, unresolved targets).
• Add intake dashboard: render the census as a clean markdown backlog page; dry-run prints to stdout, --apply writes PyAutoMind/dashboard.md.
• Wire both subcommands + --json through intake.sh.
• Update intake docs and the /intake skill body; move census/dashboard out of "planned follow-up" (repair stays planned).
• Generate and commit the first dashboard.md to PyAutoMind.

Detailed implementation plan

Affected Repositories

• PyAutoBrain (primary — agent code + docs)
• PyAutoMind (generated dashboard.md page)

Branch Survey

Repository	Current Branch	Dirty?
./PyAutoBrain	docs/build-capabilities-consumer-column (clean, 1 doc commit ahead of main)	clean
./PyAutoMind	main	clean

Suggested branch: feature/intake-census-dashboard (cut from origin/main)

Implementation Steps

1. agents/conductors/intake/_intake.py — add parse_header(text): extract the light Type:/Target:/Difficulty:/Autonomy:/Priority:/Status: header lines (intake's own convention; no YAML).
2. agents/conductors/intake/_intake.py — add census(mind): walk WORK_TYPES folders + triage/ (skip README.md; count issued/ separately as already-dispatched). Per-prompt record: path, folder-derived work-type/target, title, header fields, missing-field flags. Aggregates: counts by work-type, target, difficulty, status; hygiene list.
3. agents/conductors/intake/_intake.py — add render_dashboard(census): markdown page with summary count table, per-work-type sections with linked prompt tables (target/difficulty/autonomy/priority), triage + hygiene sections, and a "generated by intake dashboard — do not hand-edit" stamp. No Heart/health data — this is the Mind backlog view, honouring the agent's "never emit a health dashboard" rule.
4. agents/conductors/intake/_intake.py — main(): add census subcommand (always read-only) and dashboard (stdout on dry-run; writes dashboard.md under --apply); both support --json.
5. agents/conductors/intake/intake.sh — add census/dashboard to the known-subcommand list and the help header comment.
6. Docs — agents/conductors/intake/AGENTS.md modes table (census/dashboard shipped; repair stays planned), INTAKE_TAXONOMY.md short section, skills/intake/intake.md (the /intake command body), skills/COMMANDS.md routing line.
7. PyAutoMind/dashboard.md — run intake dashboard --apply against the live Mind; commit the first page via prompt_sync_push.

Key Files

• PyAutoBrain/agents/conductors/intake/_intake.py — analysis core (stdlib-only; writes only under --apply)
• PyAutoBrain/agents/conductors/intake/intake.sh — CLI front door / subcommand routing
• PyAutoBrain/agents/conductors/intake/AGENTS.md — agent contract; names census/dashboard as the planned follow-up
• PyAutoBrain/agents/faculties/sizing/_sizing.py — shared WORK_TYPES / repo sets consumed by the census walk
• PyAutoMind/dashboard.md — the generated backlog page (new)

Testing

PyAutoBrain has no pytest suite — validate via CLI: dry-run / --json / --apply against a temp Mind fixture (missing headers, triage files, empty folders), then against the live Mind; confirm dashboard.md renders cleanly on GitHub.

Original Prompt

Click to expand starting prompt

Add census and dashboard modes to the PyAutoBrain intake agent

Type: feature
Target: PyAutoBrain
Repos:

• PyAutoBrain
• PyAutoMind
  Difficulty: medium
  Autonomy: supervised
  Priority: normal
  Status: formalised

Add census and dashboard modes to the PyAutoBrain intake agent. A census mode inventories every filed PyAutoMind prompt (work-type, target, difficulty, status), and a dashboard mode renders that census as a clean summary page committed to the PyAutoMind repo. Touches the intake agent code in PyAutoBrain plus a generated page in PyAutoMind.

[Jammy2211]

Shipped in PR #31 (feature/intake-census-dashboard).

• intake census — read-only backlog inventory: 83 filed prompts, 261 already issued, 77 legacy prompts flagged headerless (hygiene, not errors). Taxonomy folder stays authoritative for work-type/target.
• intake dashboard — renders the census as the Mind backlog page; --apply wrote the first one: PyAutoMind/dashboard.md (PyAutoLabs/PyAutoMind@317feb3).
• Validated on a fixture Mind + the live Mind; classify/ideas regression-clean; --help sed-range leak fixed along the way.
• Note: shipped without pending-release — the label doesn't exist on this infra repo (no PyPI flow). Heart was YELLOW on unrelated science-stack items; acknowledged before ship.
• repair mode remains the planned follow-up.