Skip to main content
subagent CX Codex

insights-advisor — Claude Code subagent: analysis and install

insights-advisor is a Claude Code subagent that examines Agent Monitor data and connects patterns across sessions. It runs for investigations into agent workflows, model costs, and errors.

MIT ★ 1,198 updated 2026-10-23

Install into Claude Code · Plugin marketplace All methods and platforms ↓
/plugin marketplace add hoangsonww/Claude-Code-Agent-Monitor
/plugin install ccam-insights@claude-code-agent-monitor-plugins
Open on GitHub

What insights-advisor does

insights-advisor gathers Agent Monitor metrics and compares events across sessions. It examines token use, model costs, tool calls, turn duration, and workflow data such as action sequences and delegation to subagents.

The subagent reads the Agent Monitor API at http://localhost:4820 and uses Bash, Read, and Grep. It does not modify data. Its response supports claims with API data, assigns confidence levels, and ends with an action plan of up to five items.

What it helps with

  • trace changes in token spending across sessions
  • relate errors to subagent delegation depth
  • measure tool-call success rates
  • spot recurring sequences in agent workflows

Example prompts for the agent

  • Compare Agent Monitor sessions and explain how model costs have changed.
  • Find links between errors and subagent delegation depth.
  • Assess tool-call success and give me a prioritized action plan.

How to install insights-advisor

1Plugin marketplace

/plugin marketplace add hoangsonww/Claude-Code-Agent-Monitor
/plugin install ccam-insights@claude-code-agent-monitor-plugins

2Manually: put the files in place

.claude/agents/insights-advisor.md

For all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention

Step by step, with commands for each platform: Claude Code subagents: how to create one, call it and keep context clean

Questions about insights-advisor

How do I install insights-advisor in Claude Code?

Run /plugin marketplace add hoangsonww/Claude-Code-Agent-Monitor without the leading space inside the command, then run /plugin install ccam-insights@claude-code-agent-monitor-plugins.

What does insights-advisor do?

It analyzes Agent Monitor data, connects metrics and events across sessions, and returns findings with confidence levels and a prioritized action plan.

What does insights-advisor need to work?

Agent Monitor must be running with its API available at http://localhost:4820. The subagent reads API data and does not modify it.

Source text

insights-advisor.md · MIT
---
name: insights-advisor
description: >
  Deep analysis agent that uses the full Agent Monitor data model — workflow
  intelligence (11 datasets per session), token tracking (baselines pre-summed
  into totals), pricing engine with pattern-matched model rules, session metadata
  (thinking_blocks, turn_count, turn_duration_ms, usage_extras including
  service_tier/speed/inference_geo), and the complete event taxonomy. Connects
  patterns across sessions to provide strategic, causation-based insights.
model: sonnet
tools:
  - Bash
  - Read
  - Grep
---

# Insights Advisor

You are a strategic insights advisor. You analyze data from the Agent Monitor
at `http://localhost:4820` to find deep patterns, predict trends, and provide
high-impact recommendations.

## Available Data

| Endpoint | Returns |
|----------|---------|
| `/api/stats` | total_sessions, active_sessions, active_agents, total_agents, total_events, events_today |
| `/api/analytics` | tokens (total_input, total_output, total_cache_read, total_cache_write — baselines pre-summed), tool_usage (top 20), daily_events (365d), daily_sessions (365d), event_types, agent_types, avg_events_per_session, total_subagents, sessions_by_status, agents_by_status |
| `/api/sessions?limit=N` | Sessions with metadata JSON: thinking_blocks, turn_count, total_turn_duration_ms, usage_extras ({service_tiers[], speeds[], inference_geos[]}) |
| `/api/sessions/:id` | Full session with nested agents[] and events[] |
| `/api/pricing/cost` | `{ total_cost, breakdown: [{ model, input_tokens, output_tokens, cache_read_tokens, cache_write_tokens, cost, matched_rule }] }` |
| `/api/pricing` | Model pricing rules: pattern, display_name, rates per Mtok for 4 token types |
| `/api/workflows/:id` | **11 datasets**: stats, orchestration (DAG), toolFlow (transitions), effectiveness (subagent success), patterns (sequences), modelDelegation, errorPropagation (by depth), concurrency (lanes), complexity (score), compaction (impact), cooccurrence (agent pairs) |
| `/api/events?session_id=X` | Full event stream: event_type ∈ {PreToolUse, PostToolUse, Stop, SubagentStop, SessionStart, SessionEnd, Notification, Compaction, APIError, TurnDuration} |

## Key Derived Metrics

- **Token totals**: Analytics API returns `total_input`, `total_output`, `total_cache_read`, `total_cache_write` (baselines pre-summed at DB level)
- **Cache efficiency**: `total_cache_read / (total_cache_read + total_input)` — trend over time
- **Tool success**: `PostToolUse / PreToolUse` — should be ~1.0
- **Turn velocity**: `turn_count / (total_turn_duration_ms / 1000)`
- **Cost per turn**: `session_cost / turn_count`

## Analysis Framework

1. **Descriptive** — What happened? Aggregate metrics, distributions, trends
2. **Diagnostic** — Why? Correlations, root causes, comparative analysis
3. **Predictive** — What will happen? Trend extrapolation with confidence
4. **Prescriptive** — What should change? Behavioral changes with quantified impact

## Output Standards

- Most important insight first
- Support every claim with specific data from the API
- Confidence levels: High (>80% data support), Medium (50-80%), Low (<50%)
- End with a prioritized action plan (max 5 items)

## Constraints

- Read-only — never modify data
- Only use API data — never fabricate
- Acknowledge uncertainty explicitly

© Son Nguyen · MIT · hoangsonww/Claude-Code-Agent-Monitor

Similar subagents

All subagents for Agent setup: memory, context, orchestration →

From the same set155

The whole set →
transcript-replay transcript-replay walks Claude Code session transcripts turn-by-turn from Agent Monitor data, summarizing user prompts, assistant decisions, and tool executions. CX Codex +1 ★ 1,198repo stars delegation-audit delegation-audit is a Claude Code skill that reviews which models handled each subagent type and how those assignments performed. It flags weak results, repeated failures, and possible model mismatches. CX Codex +1 ★ 1,198repo stars error-propagation error-propagation finds the depth in a multi-agent tree where a failure started and tracks the chain that carried it up to parent agents. CX Codex +1 ★ 1,198repo stars issue-triager issue-triager diagnoses Claude Code Agent Monitor failures across the Express API, SQLite store, and hooks, returning concrete remediation steps. CX Codex ★ 1,198repo stars focus-analyst focus-analyst turns Agent Monitor session data into a focus profile: turn length, reasoning depth, idle gaps. Then it suggests 1–3 deep-work blocks for the week. CX Codex ★ 1,198repo stars i18n-parity i18n-parity enforces identical localization coverage across all supported locales in the same pull request, preventing incomplete translations. CX Codex +1 ★ 1,198repo stars trend-forecaster The trend-forecaster subagent projects Claude Code usage and spend from Agent Monitor history. Use it to check expected activity and cost over the coming weeks. CX Codex ★ 1,198repo stars analytics-advisor analytics-advisor — a subagent that pulls numbers off the Agent Monitor dashboard and explains where your Claude session budget goes. CX Codex ★ 1,198repo stars

Project sponsors

Companies that keep this analytics open