Team Analysis

Generate an activity report analyzing team contributions over a date range. Combines fetched GitHub data with vault context to produce a narrative report.

When to use

• "Analyze team last month", "team report for Q3", "how did the team do in April?"
• "Analyze {member} last 2 weeks" (individual focus)

Prerequisites

• Fetched team activity data in data/YYYY-MM/ (use team-activity skill if missing)
• Team roster (team.json)

Process

Step 1: Resolve the date range

Parse the request into concrete dates. Support:

Step 2: Load data

Read the relevant JSON files from data/YYYY-MM/ directories. If data is missing for the requested range, fetch it first using the team-activity skill.

Step 3: Analyze

Team-level metrics:

• Total PRs authored, merged, and reviewed
• PR distribution by repo (where is work happening?)
• Review coverage (who's reviewing whose work?)
• Issue throughput (opened vs. closed)

Individual-level metrics:

• PRs authored and merged (count and titles)
• Reviews given (count and to whom)
• Issues closed
• Notable patterns: breadth across repos, review depth, velocity trends

Qualitative observations:

• Project momentum: which projects are moving fast vs. stalled?
• Collaboration patterns: is everyone reviewing, or is it concentrated?
• Risk signals: anyone with zero reviews? Anyone with lots of open but few merged?

Step 4: Generate report

Write a markdown report:

# Team Activity Report: {date_range}

## Summary
- {N} PRs authored ({M} merged) across {R} repositories
- {X} reviews given, {Y} issues closed
- Top contributors: ...

## By Project
### {Project/Repo Name}
- Key PRs and what they accomplished
- Who worked on it

## By Member
### {Member Name}
- PRs: [list with links]
- Reviews: [count, notable ones]
- Observations: [patterns, recognition, concerns]

## Observations
- [Cross-cutting patterns, collaboration health, risk signals]

Step 5: Save and index

1. Save report to reports/ directory with descriptive filename:
   • Team: analysis-team-YYYY-MM.md
   • Individual: analysis-{login}-YYYY-MM.md
2. Update reports/index.json:

{
  "reports": [
    {
      "id": "rpt-{timestamp}",
      "type": "team|individual",
      "member": "{login or null}",
      "from": "YYYY-MM-DD",
      "to": "YYYY-MM-DD",
      "file": "analysis-....md",
      "generated_at": "ISO-8601"
    }
  ]
}

This index lets future "analyze {member}" calls pick up where the last report left off.

Rules

• Data first. If data isn't fetched for the requested range, fetch it before analyzing. Don't guess.
• No ranking or stack-ranking. Don't compare members against each other. Analyze each person's contributions on their own merits.
• Context matters. Someone with 2 PRs who shipped a complex migration is more notable than someone with 15 trivial dependency bumps. Weight quality over quantity.
• Flag what you can't see. GitHub data misses meetings, design work, mentoring, code review depth, and cross-team coordination. Note these gaps.
• Fiscal year awareness. If the team uses a non-standard fiscal year (e.g., Q1 = Jul-Sep), respect that mapping from the team roster config.