From memorylane
Generates time reports grouped by client/project from screen activity timelines, estimating block durations for timesheets and billing.
npx claudepluginhub deusxmachina-dev/memorylane --plugin memorylaneThis skill is limited to using the following tools:
Generate a time report grouped by client and project with approximate durations. The output is what you'd submit to a client or paste into a timesheet — hours per client per day, with enough detail to justify the line items.
Generates time tracking sheets, invoice calculations from billable hours, and project profitability analytics for freelancers managing multiple clients.
Guides Next.js Cache Components and Partial Prerendering (PPR) with cacheComponents enabled. Implements 'use cache', cacheLife(), cacheTag(), revalidateTag(), static/dynamic optimization, and cache debugging.
Guides building MCP servers enabling LLMs to interact with external services via tools. Covers best practices, TypeScript/Node (MCP SDK), Python (FastMCP).
Share bugs, ideas, or general feedback.
Generate a time report grouped by client and project with approximate durations. The output is what you'd submit to a client or paste into a timesheet — hours per client per day, with enough detail to justify the line items.
Before fetching any data, ask the user two things:
Store both answers — you'll use the client list to classify time blocks and the range to set startTime/endTime.
For the requested time range, iterate one day at a time to preserve sequential context:
browse_timeline(startTime="<day start>", endTime="<day end>", limit=200, sampling="uniform")
If the response header shows results were sampled (i.e., more entries exist), re-fetch that day with a higher limit (up to 1000).
For a single-day report, one call is enough. For multi-day ranges, iterate day by day and process each batch before moving on.
Walk through each day's activities chronologically:
Map each time block to a client from the user's list:
get_activity_details(ids) on 1–2 representative entries to inspect OCR text. Keep this to genuinely ambiguous cases.If the user skipped providing a client list in Step 1 (scan-first mode), propose a client list based on the distinct projects/apps you see, ask the user to confirm or adjust, then re-classify.
Format as a markdown table, grouped by client:
## Time Report — [Date range]
### [Client A]
| Date | Time Range | Project / Task | Duration | Details |
|---|---|---|---|---|
| Mon Feb 23 | 9:00 AM – 10:15 AM | API integration | 1h 15m | Editing api.ts, testing endpoints |
| Mon Feb 23 | 2:00 PM – 3:30 PM | API integration | 1h 30m | Debugging auth flow |
| Tue Feb 24 | 10:00 AM – 11:45 AM | Dashboard | 1h 45m | Building chart components |
**Subtotal**: 4h 30m
### [Client B]
| Date | Time Range | Project / Task | Duration | Details |
|---|---|---|---|---|
| Mon Feb 23 | 11:00 AM – 12:00 PM | Landing page | 1h 0m | Designing hero section |
**Subtotal**: 1h 0m
### Internal / Unassigned
| Date | Time Range | Activity | Duration | Details |
|---|---|---|---|---|
| Mon Feb 23 | 10:15 AM – 10:30 AM | *Break* | 15m | |
| Tue Feb 24 | 9:00 AM – 9:45 AM | Email + Slack | 45m | |
**Subtotal**: 1h 0m
---
**Total tracked time**: 6h 30m
**Breaks**: 15m
**Coverage**: Mon Feb 23 – Tue Feb 24
If the range is a single day, drop the Date column. If the report period is still in progress, note it's partial.
If the user asks follow-up questions after the report:
search_context(query) for client-specific or project-specific queries (e.g., "how long did I spend on Client A's API this week?").get_activity_details(ids) when they want exact on-screen text from a specific time block."today" resolves to midnight local time on the server.