nx-scorecard

name: nx-scorecard description: > Compute Nx CLI scorecard metrics for a month: GitHub issues/PRs (nrwl/nx), Linear high-priority misc issues, and npm download counts with YoY growth. Outputs a Notion-ready markdown table. Use when user says "scorecard", "nx scorecard", "cli scorecard", "monthly scorecard", or asks for the Nx CLI team's monthly issue/PR/download metrics. allowed-tools: - Bash - WebFetch - mcplinear-serverlist_issues - mcplinear-serverlist_teams

Nx CLI Scorecard

Compute all scorecard metrics for the Nx CLI team for the current month (or the month specified in the user's request / $ARGUMENTS).

Determine Target Month

If the request contains a month (e.g., "March", "2026-03", "Feb 2026"), use that month.
Otherwise, use the current month based on today's date.
Derive: YEAR, MONTH_NUM (zero-padded), MONTH_NAME, FIRST_DAY (YYYY-MM-01), LAST_DAY (YYYY-MM-last day of month).
For partial months (current month not yet ended), use today's date as the end date for closed issues.

GitHub Auth (no `gh` CLI)

The gh CLI is banned on this machine. Use curl against https://api.github.com with a bearer token.

Get the token from 1Password (invoke the op-request-reason skill first, since op triggers a 1P prompt):

export GITHUB_TOKEN="$(op read 'op://Employee/API Keys/github_token')"

Reusable curl helper for the GitHub search API (returns total_count):

gh_count() {  # $1 = url-encoded q value
  curl -sS --fail \
    -H "Authorization: Bearer $GITHUB_TOKEN" \
    -H "Accept: application/vnd.github+json" \
    -H "X-GitHub-Api-Version: 2022-11-28" \
    "https://api.github.com/search/issues?q=$1&per_page=1" | jq '.total_count'
}

GitHub's search q must be URL-encoded (spaces -> + is fine in a query, but : and quotes in label filters need encoding). Build each query string and pass it to gh_count.

Data Collection

Run queries in parallel where possible.

1. GitHub Issues (nrwl/nx)

All use is:issue. Replace {FIRST_DAY}/{LAST_DAY} with the target month bounds.

Open issues (snapshot): repo:nrwl/nx is:issue state:open created:2017-09-01..{LAST_DAY}
Open issues created this month: repo:nrwl/nx is:issue state:open created:{FIRST_DAY}..{LAST_DAY}
Open high-priority (snapshot): repo:nrwl/nx is:issue state:open created:2017-09-01..{LAST_DAY} label:"priority: high"
Open high-priority created this month: repo:nrwl/nx is:issue state:open created:{FIRST_DAY}..{LAST_DAY} label:"priority: high"
Closed issues (during month): repo:nrwl/nx is:issue state:closed closed:{FIRST_DAY}..{LAST_DAY}
Closed high-priority (during month): repo:nrwl/nx is:issue state:closed closed:{FIRST_DAY}..{LAST_DAY} label:"priority: high"

2. GitHub PRs (nrwl/nx)

Count only non-draft PRs (draft:false).

Open non-draft PRs (snapshot): repo:nrwl/nx is:pr state:open draft:false
Open non-draft PRs created this month: repo:nrwl/nx is:pr state:open draft:false created:{FIRST_DAY}..{LAST_DAY}

3. Linear High-Priority Misc Issues

Nx CLI team issues in the "Miscellaneous" project OR no project, with Urgent (priority=1) or High (priority=2) priority, in an active state (not Done/Cancelled).

Strategy: Query ALL open Nx CLI urgent+high issues, then filter in post-processing. More reliable than querying "no project" directly.

Confirm the team first (mcp__linear-server__list_teams -> "Nx CLI"). Then run these 4 queries in parallel via mcp__linear-server__list_issues:

team: "Nx CLI", priority: 1, state: "started", limit: 250
team: "Nx CLI", priority: 2, state: "started", limit: 250
team: "Nx CLI", priority: 1, state: "unstarted", limit: 250
team: "Nx CLI", priority: 2, state: "unstarted", limit: 250

From the combined results, keep only issues where project is "Miscellaneous" OR project is null/empty/missing. Deduplicate by issue id and count unique issues.

4. npm Downloads

# Current month
curl -s "https://api.npmjs.org/downloads/point/{FIRST_DAY}:{LAST_DAY}/nx" | jq '.downloads'
# Same month last year (YoY)
curl -s "https://api.npmjs.org/downloads/point/{LAST_YEAR_FIRST_DAY}:{LAST_YEAR_LAST_DAY}/nx" | jq '.downloads'

YoY growth: (this_month - same_month_last_year) / same_month_last_year * 100, rounded to nearest whole percent. If the current month is incomplete, note the YoY comparison is partial.

Output Format

## Nx CLI Scorecard — {MONTH_NAME} {YEAR}

| Metric | Value |
|--------|-------|
| Open issues | {total} ({this_month} this month) |
| Open high-priority issues | {total} ({this_month} this month) |
| Closed issues | {count} |
| Closed high-priority issues | {count} |
| Open Linear high-priority misc issues | {count} |
| Non-draft PRs (open) | {total} ({this_month} this month) |
| Total monthly downloads | {formatted, e.g. 36.0M} |
| YoY growth (monthly downloads) | {percent}% |

Format downloads in millions with one decimal (e.g., 33.2M, 36.0M).

If the month is not yet complete, add a note: *Data as of {TODAY}. Month in progress.*

Important Notes

GitHub search API has rate limits. On a rate-limit response, wait and retry once.
"Open issues" snapshot: if running mid-month, use today's date as the upper bound instead of LAST_DAY.
The creation floor 2017-09-01 filters out ancient pre-active issues.
Do NOT include PRs in issue counts (is:issue in all issue queries).