$npx -y skills add microsoft/What-I-Did-Copilot --skill whatididGenerate a daily analytics report of what GitHub Copilot helped accomplish. Shows tasks completed, human effort equivalent, token usage, and a narrative story. Sends a formatted email summary. Use when the user asks about their daily activity, what Copilot helped with today, or w
| 1 | # What I Did (Copilot) — Impact Report Generator |
| 2 | |
| 3 | Run the following from the repo root to generate and email today's activity report. Use `--email` with **no argument** so the tool auto-detects the user's address (from `gh` / git config) — never substitute a literal placeholder like `user@example.com`: |
| 4 | |
| 5 | ```bash |
| 6 | python whatidid.py --email |
| 7 | ``` |
| 8 | |
| 9 | If the user asks for a specific date range, use: |
| 10 | ```bash |
| 11 | python whatidid.py --date 30D --email |
| 12 | ``` |
| 13 | |
| 14 | To send to a specific address instead of the auto-detected one, pass it explicitly: `--email someone@example.com`. |
| 15 | |
| 16 | After running, tell the user: |
| 17 | - How many sessions and projects were found |
| 18 | - The headline and primary focus identified |
| 19 | - The total human effort estimate vs elapsed time (leverage ratio) |
| 20 | - That the email has been sent (or HTML saved) |
| 21 | |
| 22 | The report is always opened in the browser automatically — no need to add --html. |
| 23 | |
| 24 | If there are no sessions for the date, explain that Copilot session data is stored in the user's Copilot session-state directory (for example, `~/.copilot/session-state/` on macOS/Linux or `%USERPROFILE%\.copilot\session-state\` on Windows) and suggest checking the date. |
| 25 | |
| 26 | For full methodology, see [effort-estimation-methodology.md](../../docs/effort-estimation-methodology.md). |