| name | reporting |
| description | The Code reporting discipline AND the reporting-system operations command. Discipline half — the "one spine" rule (every reportable unit of work lands one Weekly Work Tracker row, Source-tagged), field mapping, Bridge auto-mirror tie-in. Operations half (proposal 07) — /reporting status (registry tail, trust metrics, drift, next scheduled runs, log locations), /reporting run <pipeline>, /reporting heal. Use when finishing reportable work, when asked "how should I report this" / "log this for the report", at session end to capture unreported work, or for pipeline ops — "/reporting", "reporting status", "is reporting healthy", "run the health check", "why is a report row missing/stale". |
Reporting Discipline (Claude Code)
The whole system is "capture once, report many." Reporting quality is a capture problem — the
renderers (team dashboard + sponsor site) are already good. Your job here is to make sure every
reportable unit of work lands in the one spine: the AI Enablement Weekly Work Tracker.
See the full system map + improvement proposal: docs/reporting/reporting-system-overview.md.
The one rule
After any reportable unit of work, ensure exactly one Weekly Work Tracker row exists for it,
with Source = Code.
"Reportable" = a Ship · Fix · Feature · Integration · Deploy · Decision · Research · Design unit a
sponsor or the team would want to see. Not every edit — the shippable/decision-worthy ones.
Two ways a row gets created (don't double-log)
| Situation | Path | You do |
|---|
| Work done as a Bridge writeback to Eve/Donna | notion-bridge auto-mirror (step 3b) creates the row | Follow the Bridge writeback protocol — the row is automatic. Verify it landed. |
| Work not going through a Bridge writeback (most CCv3 work) | /log-work skill | Invoke /log-work (or the log-work field call) to create the row |
Dedup: both paths dedup by Writeback URL. If a Bridge writeback already mirrored a row, do not
also /log-work the same item.
Field mapping (every row)
| Field | Value |
|---|
| Title | verb-led, specific (e.g. "Wired interactive 3-view reporting toggle") |
| Source | Code (autonomous Code work) — Dave/Team only if attributing to a human |
| Type | Ship · Fix · Feature · Integration · Deploy · Decision · Research · Design · Meeting |
| Impact | High · Mid · Low (default Mid) |
| Date | the work date (today) |
| Dashboard Section | drives which report section it surfaces in (table below) |
| Project | relation to the relevant FourthOS Project, when one applies |
| Commits | SHA(s) if code shipped |
| Writeback URL | the Bridge child page / PR / artifact URL, if any |
| Published in Report | leave unchecked — the collect_weekly_work_tracker collector sets it |
Dashboard Section decision table
Highlight (headline win) · Decision (made/needed) · Deep Dive (named project progress) ·
Integration (connected systems) · Metric (a number) · Adoption (usage/rollout) ·
Roadmap (forward plan) · None (logged, not for the report).
The one-spine rule includes FourthOS work
The scheduled fourthos-weekly sponsor run writes the Notion cockpit but historically did not log
to the Tracker. Per the "one spine, two cuts" model: any reportable FourthOS work should also land a
Tracker row (Source=Code, Dashboard Section=Highlight) so it appears in the team report too. If
you ship sponsor-facing work outside the scheduled run, log it here.
Refresh the living card too. After reportable work on a FourthOS project that has a living status
card, run /project-card refresh "<project>" so its Notion card reflects the change (the daily
CCv3-Project-Cards sweep will catch it otherwise, but refresh now if it's report-worthy). The card is
a machine mirror — you change the source data (Projects DB / Decisions & Outputs), the card regenerates;
never hand-edit the card embed. See the project-card skill.
How this feeds reporting
- Team dashboard (
/weekly-report): collect_weekly_work_tracker reads unpublished Tracker rows,
groups them by Dashboard Section, and marks them published. Your rows ARE the report.
- Sponsor site (
fourthos-weekly): renders the VP/CTO cut; as the system unifies, the Tracker's
"what shipped" signal feeds the sponsor "What moved" lane.
- Prereq: the
ai-reporting MCP must be registered for /weekly-report to run (see overview §3 —
currently unregistered; flag it if the collector errors).
Session-end habit
Before ending a substantive session, ask: "What did I ship/decide that isn't yet a Tracker row?" —
then log it. Unlogged work is invisible to every report.
Reporting Operations (/reporting status | run | heal) — proposal 07
The 60-second answer to "is the reporting system healthy?" plus the common operations, as
muscle memory. All commands are Windows-safe; ntn calls in scripts already handle the
</dev/null contract internally. Deep background: scripts/report-registry/README.md
(schema, conventions, health strip, trust metrics) and the field-guide artifact.
/reporting status — the 60-second check
Run these (all read-only; safe anytime):
node scripts/report-registry/trust-metrics.mjs
node scripts/report-registry/check-drift.mjs
node scripts/report-registry/watchdog.mjs --dry-run
Then, as needed:
- Registry tail: newest rows per type are in the check-drift/watchdog output; raw query via
ntn datasources query c7d2d9e3-d388-4640-a66e-88f7dd50f854 (TSV, not JSON).
- Next scheduled runs:
schtasks /query /tn "<task>" for the tasks below, or PowerShell
Get-ScheduledTaskInfo <task>.
- Dated logs:
~/.claude/logs/<pipeline>/<YYYY-MM-DD>.log — pipelines: health-check/,
fourthos-weekly/, dashboard-sync/, self-improvement/ (project-cards logs to
scripts/project-cards/sweep.jsonl; VP Weekly to ai-report-card/ run output).
- Hub surfaces: the Reports hub launcher + Reporting Health strip are machine-owned and
refreshed by
refresh-pages.mjs — if they lag the registry, that IS the finding (see heal).
/reporting run <pipeline> — manual invocation
| Pipeline (registry type) | Scheduled task | Manual command |
|---|
| System Health | CCv3-Health-Check (Fri 08:00) | scripts\scheduled-health-check.bat (Notion mirror = node scripts/report-registry/health-mirror.mjs, ntn-only) |
| FourthOS Sponsor | CCv3-FourthOS-Weekly (Thu 07:00) | scripts\fourthos-weekly\scheduled-fourthos-weekly.bat |
| VP Weekly | AIWeeklyReport (Thu 06:00 — needs logged-in session until the IT batch-logon ticket clears) | scripts\run-ai-weekly-report.bat |
| Team Dashboard | CCv3-Dashboard-Sync (daily) | scripts\sync-tasks-dashboard.ps1 (LLM path; deterministic replacement is BLOCKED — see the 2026-07-17 efficiency handoff GAP-1..4) |
| Project Portfolio (cards) | CCv3-Project-Cards (daily) | node scripts/project-cards/sweep.mjs (--dry-run is safe: no publishes, no claude spawn) |
| Self-Improvement | CCv3-Self-Improvement-Research (daily 07:30) | headless research loop; digests queue in docs/self-improvement/PENDING-DIGEST.md for interactive Bridge sweep |
Every run must land a registry row (make-run + upsert are wired into the wrappers); the upsert
auto-refreshes that type's child page + hub launcher (proposal 02).
/reporting heal — when something is wrong
| Symptom | Action |
|---|
| Expected period has no row (missed run) | node scripts/report-registry/watchdog.mjs (no flag) writes the visible watchdog: miss row; then re-run the pipeline manually and let the genuine row supersede it |
| Hub/child pages lag the registry | node scripts/report-registry/refresh-pages.mjs (all) or --type "<Type>" (scoped) |
| Health page mirror stale | node scripts/report-registry/health-mirror.mjs (--dry-run first to preview sections) |
| History gap (rows never emitted) | node scripts/report-registry/backfill.mjs --dry-run then without the flag — backfill rows carry the backfill: prefix and never masquerade as current |
| Wrapper exits 0 but no artifact | That's the closed 11-week class — check the dated log for permission-denied vs mcp-unavailable (rule headless-claude-mcp.md); fix the grant, never reclassify as SKIP |
A .ps1/.bat wrapper change | Test under real PS 5.1 (powershell.exe -File), never only pwsh 7; no Tee-Object -Encoding; set vars ABOVE parenthesized batch blocks |
Conventions that keep this trustworthy (enforced in make-run.mjs, proposal 08): corrective rows
carry backfill:/remap: prefixes; artifact URLs must be durable http(s); one period format per
cadence (VP Weekly ISO-week is the lone exception); every type has a watchdog schedule entry.
Related
/log-work — the row-creation primitive (field parsing, Source detection, dedup).notion-bridge — the auto-mirror (step 3b) + Eve/Donna handoff queues.weekly-report — the team dashboard half (Code data + Cowork/Desktop editorial).fourthos-weekly — the scheduled sponsor site.docs/reporting/reporting-system-overview.md — full map + "one spine, two cuts" proposal.docs/reporting/donna-reporting-addendum.md — Donna's parallel rule (Source=Donna).
