nntrivi2001
57dc91585d
Internal SmartGift build of a Claude Code monitoring dashboard. Lanes: a durable unit of parallel agent work, one per working directory, tracked across session restarts. Managed lanes are git worktrees the dashboard provisions and can reset or remove behind a three-check destroy guard and a counted preflight; adopted lanes are directories you already own and are never destroyable. Pipelines: a lane moves through pipeline stages. A stage the agent declares with evidence renders green; a stage inferred from the tool-event stream renders dashed amber and never counts as done. Detection is forward-only within a 30-minute window, and never writes the declared stage. Workspace: one page at /run with a lane grid, the selected lane's pipeline, and a full Claude console behind a disclosure.
3.3 KiB
description
| description |
|---|
| Compile a month-over-month retrospective from Agent Monitor data — sessions, cost, token volumes, completion rate, top projects by working directory, and notable shifts versus the prior month. Uses daily_sessions/daily_events (365d) from analytics, the session list, and the pricing cost breakdown. Use when doing a monthly retrospective or planning the month ahead. |
Monthly Review
Generate a month-over-month productivity retrospective from Agent Monitor data.
Input
The user provides: $ARGUMENTS
This may be:
- "this month" or empty (default: the current calendar month to date)
- "last month" for the previous full calendar month
- A specific month: "2026-02" or "February 2026"
The comparison period is always the immediately preceding calendar month.
Data Sources
| Endpoint | Returns |
|---|---|
GET /api/analytics |
daily_sessions and daily_events (365d) for monthly bucketing and trends; tokens (total_input/output/cache_read/cache_write — baselines pre-summed); tool_usage (top 20); sessions_by_status |
GET /api/sessions?limit=500 |
Sessions with started_at, ended_at, status, model, cwd, cost, and metadata (turn_count, thinking_blocks) — for per-project (cwd) grouping and completion rate |
GET /api/pricing/cost |
total_cost and per-model breakdown (input/output/cache tokens, cost, matched_rule) |
Report Sections
1. Month at a Glance
Compare the target month to the prior month in a table:
| Metric | This Month | Last Month | Change |
|---|---|---|---|
| Sessions | N | N | ▲/▼ N% |
| Total Cost | $X.XXXX | $X.XXXX | ▲/▼ N% |
| Tokens (in/out/cache) | N | N | ▲/▼ N% |
| Completion Rate | N% | N% | ▲/▼ N pts |
| Active Days | N | N | ▲/▼ |
Derive monthly buckets from daily_sessions / daily_events. Completion rate =
completed sessions / total sessions for the month (from sessions_by_status and
the filtered session list).
2. Top Projects (by cwd)
Group the month's sessions by cwd. For the top 5–8 projects, list session count,
total cost, completion rate, and dominant model. Note any project that newly
appeared or dropped off versus last month.
3. Cost & Token Breakdown
From /api/pricing/cost, show cost per model and the dominant token type. Compute
cache hit rate = total_cache_read / (total_cache_read + total_input) and compare
to last month. Currency to 4 decimals.
4. Tool & Workflow Shifts
From tool_usage, highlight the tools that rose or fell most month-over-month, and
any new tool adopted. Flag rising error/Compaction activity if present.
5. Notable Shifts & Narrative
Three to five plain-language observations: what changed, why it likely changed, and what it implies (e.g., "cost up 22% but sessions flat → heavier per-session work").
6. Focus for Next Month
Two to four prioritized, actionable goals grounded in the numbers above.
Output
- Markdown report with emoji-light, scannable section headers.
- Tables for all month-over-month comparisons; ▲ / ▼ for deltas.
- Currency in USD to 4 decimals; tokens with thousands separators.
- Lead with a 2–3 sentence executive summary, then the sections in order.
- Cite only numbers returned by the API; if a month has no data, say so explicitly.