Productivity Score

Turns Agent Monitor's raw telemetry into a single 0-100 score by weighting five dimensions: completion rate, cache efficiency, tool success ratio, turn velocity, and cost per session. It pulls from four API endpoints, calculates metrics like cache hit rate and turn speed, then gives you a letter grade with ASCII bar charts and actionable improvements. Honestly most useful when you're comparing time periods or trying to justify why you spent three hours optimizing prompt caching. The scoring weights are opinionated but reasonable, and seeing "your turns are slow because you're using Opus for everything" hits harder than staring at raw JSON. Run it weekly or after major workflow changes to catch regressions before they become expensive habits.

GitHub stars: 455
last updated: Jun 11, 2026

View on GitHub

Install

$ npx -y skills add hoangsonww/claude-code-agent-monitor --skill productivity-score --agent claude-code

Installs into .claude/skills of the current project.

Files

Productivity Score

Calculate a productivity scorecard from the Agent Monitor's real data.

Input

The user provides: $ARGUMENTS

Options: "today", "this week", "last 30 days", a session ID, or "compare" for period comparison.

Data Sources

Endpoint	Returns
`GET /api/analytics`	Token totals (`total_input`, `total_output`, `total_cache_read`, `total_cache_write` — baselines pre-summed), tool_usage top 20, daily_events/sessions, event_types, sessions_by_status, agents_by_status, avg_events_per_session, total_subagents
`GET /api/sessions?limit=100`	Sessions with metadata JSON: `thinking_blocks`, `turn_count`, `total_turn_duration_ms`, `usage_extras` (service_tier, speed, inference_geo)
`GET /api/pricing/cost`	Total cost with per-model breakdown
`GET /api/workflows/{sessionId}`	11 workflow datasets: stats, orchestration, toolFlow, effectiveness, patterns, modelDelegation, errorPropagation, concurrency, complexity, compaction, cooccurrence