Commit Insights Skill

Trigger Phrases

Say any of these to activate this skill:

"commit insights" "engineering metrics" "team velocity" "pr cycle time" "code churn" "engineering health" "team stats"

tool_search tool_install canvas_create canvas_update canvas_api_schema canvas_api_seed memory_read memory_write send_message web

Analyze git activity and produce engineering health metrics:

1. Connect — Check if GitHub integration is installed via `tool_search`. If not:

• `tool_install({ name: "github" })` to connect via Composio OAuth

2. Configure — Read tracked repos and team from memory (key: `git_insights_config`)

• Repos to analyze, team member GitHub usernames

• If not configured, ask the user which repos/team to track

3. Fetch data — For each configured repo:

• `GITHUB_LIST_COMMITS` — all commits in the analysis window (7 days default)

• `GITHUB_LIST_PULL_REQUESTS` — open, merged, and closed PRs

• PR details: time from open to first review, time to merge, review rounds

4. Compute metrics —

• **Weekly commits:** total and per-developer breakdown

• **PR cycle time:** median time from PR open to merge

• **Time to first review:** median time from PR open to first review comment

• **Code churn:** files changed frequently (>3 times in a week)

• **PR aging:** open PRs sorted by age, flagging those >3 days without review

• **Top contributors:** ranked by commits + reviews

5. Build canvas — Create or update the insights dashboard:

• KPIs: weekly commits, avg PR cycle time, top contributor, active PRs

• Team leaderboard table (developer, commits, PRs merged, reviews given)

• PR aging table (PR title, author, age, status, reviewers)

• Code churn hotspots (files changed most frequently)

6. Weekly report — On weekly heartbeat:

• Compile full engineering health report

• Compare to previous week (trending up/down)

• Post to configured channel via `send_message`

7. Persist — Save weekly snapshots to memory for trend analysis