olakai-reports（奥拉凯报道）

Olakai Analytics Reports

This skill enables terminal-based analytics report generation using the Olakai CLI, eliminating the need to access the web UI for analytics insights.

For full documentation, see: https://app.olakai.ai/llms.txt

Prerequisites

Before generating reports, ensure:

# 1. CLI is authenticated
olakai whoami

# 2. You have the agent ID (if reporting on specific agent)
olakai agents list --json | jq '.[] | {id, name}'

Report Generation Workflow

1. Gather context - Determine agent ID, date range, and report type
2. Query data - Use CLI commands with --json flag
3. Process output - Extract relevant metrics using jq
4. Generate visualizations - Create ASCII charts and markdown tables
5. Present report - Format and display the complete report

Available Data Sources

Report Type 1: Usage Summary Report

Shows total usage metrics across events, tokens, models, and agents.

Data Collection

# Get recent events with analytics
olakai activity list --limit 100 --include-analytics --json > /tmp/events.json

# Extract summary metrics
cat /tmp/events.json | jq '{
  total_events: (.prompts | length),
  total_tokens: ([.prompts[].tokens // 0] | add),
  avg_tokens: ([.prompts[].tokens // 0] | add / length | floor),
  unique_models: ([.prompts[].model] | unique | length),
  models: ([.prompts[].model] | group_by(.) | map({model: .[0], count: length})),
  unique_agents: ([.prompts[].app] | unique | length),
  agents: ([.prompts[].app] | group_by(.) | map({agent: .[0], count: length})),
  success_rate: (([.prompts[] | select(.status != "error")] | length) / (.prompts | length) * 100 | floor)
}'

Report Template

# Usage Summary Report
Generated: [DATE]
Period: Last [N] events

## Overview
| Metric | Value |
|--------|-------|
| Total Events | [COUNT] |
| Total Tokens | [TOKENS] |
| Avg Tokens/Event | [AVG] |
| Success Rate | [RATE]% |

## Events by Model
[ASCII BAR CHART]

## Events by Agent
[ASCII BAR CHART]

Example Output

# Usage Summary Report
Generated: 2025-01-21
Period: Last 100 events

## Overview
| Metric | Value |
|--------|-------|
| Total Events | 100 |
| Total Tokens | 45,230 |
| Avg Tokens/Event | 452 |
| Success Rate | 98% |

## Events by Model
gpt-4o          ████████████████████████████████████ 45
gpt-4o-mini     ██████████████████████ 28
claude-3-5      ████████████████ 20
gpt-3.5-turbo   █████ 7

## Events by Agent
code-assistant  ████████████████████████████████ 40
data-analyzer   ████████████████████████ 30
chat-support    ████████████████████ 25
test-agent      ████ 5

Report Type 2: KPI Trends Report

Shows KPI values over time with period-over-period comparisons.

Data Collection

# Get KPIs with daily breakdown
olakai activity kpis --period daily --json > /tmp/kpis_daily.json

# Get KPIs with weekly breakdown
olakai activity kpis --period weekly --json > /tmp/kpis_weekly.json

# Extract trend data
cat /tmp/kpis_daily.json | jq '{
  period: "daily",
  kpis: [.kpis[] | {
    name: .name,
    current: .value,
    trend: .trend,
    breakdown: .breakdown
  }]
}'

For Custom KPIs with Agent Filter

# Get custom KPIs for specific agent
olakai activity kpis --agent-id AGENT_ID --period daily --json | jq '.kpis'

# List KPI definitions
olakai kpis list --agent-id AGENT_ID --json | jq '.[] | {name, unit, aggregation}'

Report Template

# KPI Trends Report
Generated: [DATE]
Agent: [AGENT_NAME] (or "All Agents")
Period: [PERIOD]

## Raw Metrics
| Metric | Current | Previous | Change |
|--------|---------|----------|--------|
| Interaction Volume | [VAL] | [PREV] | [+/-]% |
| Token Consumption | [VAL] | [PREV] | [+/-]% |

## Metric Slot KPIs
| KPI | Current | Previous | Change |
|-----|---------|----------|--------|
| Execution Cost (USD) | $[VAL] | $[PREV] | [+/-]% |
| Time Saved (min) | [VAL] | [PREV] | [+/-]% |
| Value Created (USD) | $[VAL] | $[PREV] | [+/-]% |
| Governance Compliance (%) | [VAL]% | [PREV]% | [+/-]% |

## Composites
| KPI | Current | Previous | Change |
|-----|---------|----------|--------|
| ROI | [VAL]x | [PREV]x | [+/-]% |

## Custom KPIs
| KPI | Value | Unit | Aggregation |
|-----|-------|------|-------------|
| [NAME] | [VAL] | [UNIT] | [AGG] |

## Daily Trend (Last 7 Days)
[ASCII LINE CHART]

Example Output

# KPI Trends Report
Generated: 2025-01-21
Agent: code-assistant
Period: Last 7 days

## Raw Metrics & Slot KPIs
| KPI | Current | Previous | Change |
|-----|---------|----------|--------|
| Interaction Volume | 847 | 792 | +7% |
| Execution Cost | $127.50 | $118.80 | +7% |
| Time Saved | 1,694 min | 1,584 min | +7% |
| Value Created | $4,235 | $3,960 | +7% |
| Governance Compliance | 99.2% | 98.5% | +0.7% |
| ROI | 33.2x | 33.3x | -0.3% |

## Custom KPIs
| KPI | Value | Unit | Aggregation |
|-----|-------|------|-------------|
| Code Reviews | 156 | count | SUM |
| Bugs Found | 23 | count | SUM |
| Avg Response Quality | 4.7 | score | AVERAGE |

## Daily Interactions (Last 7 Days)
     150 ┤                           ╭──
     125 ┤              ╭────────────╯
     100 ┤    ╭─────────╯
      75 ┤────╯
      50 ┤
         └──────────────────────────────
          Mon  Tue  Wed  Thu  Fri  Sat  Sun

Report Type 3: Risk Analysis Report

Shows risk distribution, blocked events, and sensitivity patterns.

Data Collection

# Get events with risk data
olakai activity list --limit 200 --include-analytics --json > /tmp/events.json

# Extract risk metrics
cat /tmp/events.json | jq '{
  total_events: (.prompts | length),
  high_risk: ([.prompts[] | select(.riskScore >= 7)] | length),
  medium_risk: ([.prompts[] | select(.riskScore >= 4 and .riskScore < 7)] | length),
  low_risk: ([.prompts[] | select(.riskScore < 4)] | length),
  blocked: ([.prompts[] | select(.status == "blocked")] | length),
  blocked_percentage: (([.prompts[] | select(.status == "blocked")] | length) / (.prompts | length) * 100),
  sensitivity_labels: ([.prompts[].sensitivityLabel] | group_by(.) | map({label: .[0], count: length})),
  avg_risk_score: ([.prompts[].riskScore // 0] | add / length)
}'

Report Template

# Risk Analysis Report
Generated: [DATE]
Period: Last [N] events

## Risk Overview
| Metric | Value |
|--------|-------|
| Total Events Analyzed | [COUNT] |
| High Risk Events | [COUNT] ([%]%) |
| Blocked Events | [COUNT] ([%]%) |
| Average Risk Score | [SCORE]/10 |

## Risk Distribution
[ASCII BAR CHART]

## Events by Sensitivity Label
[ASCII BAR CHART]

## High-Risk Event Details (Recent)
| Time | Agent | Risk Score | Reason |
|------|-------|------------|--------|
| [TIME] | [AGENT] | [SCORE] | [REASON] |

Example Output

# Risk Analysis Report
Generated: 2025-01-21
Period: Last 200 events

## Risk Overview
| Metric | Value |
|--------|-------|
| Total Events Analyzed | 200 |
| High Risk Events | 8 (4%) |
| Blocked Events | 3 (1.5%) |
| Average Risk Score | 2.3/10 |

## Risk Distribution
Low (0-3)     ████████████████████████████████████████ 172 (86%)
Medium (4-6)  ████████ 20 (10%)
High (7-10)   ████ 8 (4%)

## Events by Sensitivity Label
Public        ████████████████████████████████████ 145
Internal      ██████████████████ 42
Confidential  ████ 10
Restricted    █ 3

## High-Risk Events (Recent 5)
| Time | Agent | Score | Model |
|------|-------|-------|-------|
| 10:23 | data-export | 8.5 | gpt-4o |
| 09:15 | chat-support | 7.2 | gpt-4o |
| 08:42 | code-assist | 7.0 | claude-3-5 |

Report Type 4: ROI/Efficiency Report

Shows time saved, cost metrics, and productivity gains.

Data Collection

# Get KPIs (includes ROI data)
olakai activity kpis --json > /tmp/kpis.json

# Get events with time saved data
olakai activity list --limit 100 --include-analytics --json > /tmp/events.json

# Extract efficiency metrics
cat /tmp/events.json | jq '{
  total_events: (.prompts | length),
  total_time_saved_minutes: ([.prompts[].timeSavedMinutes // 0] | add),
  avg_time_saved: ([.prompts[].timeSavedMinutes // 0] | add / length),
  total_tokens: ([.prompts[].tokens // 0] | add),
  by_task: ([.prompts[] | select(.task != null)] | group_by(.task) | map({
    task: .[0].task,
    count: length,
    time_saved: ([.[].timeSavedMinutes // 0] | add)
  }))
}'

# Get ROI from KPIs
cat /tmp/kpis.json | jq '.kpis[] | select(.name | contains("ROI") or contains("Compliance"))'

Report Template

# ROI/Efficiency Report
Generated: [DATE]
Period: Last [N] events

## Efficiency Summary
| Metric | Value |
|--------|-------|
| Total Events | [COUNT] |
| Total Time Saved | [HOURS] hours |
| Avg Time Saved/Event | [MIN] minutes |
| Estimated Cost Savings | $[AMOUNT] |

## Governance Compliance
| Metric | Value |
|--------|-------|
| Compliance Rate | [RATE]% |
| Policy Violations | [COUNT] |
| Auto-Blocked | [COUNT] |

## Time Saved by Task Type
[ASCII BAR CHART]

## ROI Breakdown
[ASCII PIE CHART or TABLE]

Example Output

# ROI/Efficiency Report
Generated: 2025-01-21
Period: Last 100 events

## Efficiency Summary
| Metric | Value |
|--------|-------|
| Total Events | 100 |
| Total Time Saved | 12.5 hours |
| Avg Time Saved/Event | 7.5 minutes |
| Estimated Cost Savings | $1,875 |

## Governance Compliance
| Metric | Value |
|--------|-------|
| Compliance Rate | 99.2% |
| Policy Violations | 2 |
| Auto-Blocked | 1 |

## Time Saved by Task Type
Code Review      ████████████████████████████████ 4.2 hrs
Bug Analysis     ██████████████████████████ 3.5 hrs
Documentation    ████████████████████ 2.7 hrs
Refactoring      ████████████████ 2.1 hrs

## Productivity Multiplier
Based on avg 7.5 min saved per interaction:
- Daily (50 events): 6.25 hours saved
- Weekly (250 events): 31.25 hours saved
- Monthly (1000 events): 125 hours saved

Report Type 5: Agent Comparison Report

Side-by-side comparison of metrics across multiple agents.

Data Collection

# Get all agents
olakai agents list --json > /tmp/agents.json

# Get events for comparison
olakai activity list --limit 500 --include-analytics --json > /tmp/events.json

# Extract per-agent metrics
cat /tmp/events.json | jq '{
  agents: ([.prompts[].app] | unique | map(. as $agent | {
    name: $agent,
    events: ([($parent.prompts // [])[] | select(.app == $agent)] | length),
    tokens: ([($parent.prompts // [])[] | select(.app == $agent) | .tokens // 0] | add),
    avg_risk: ([($parent.prompts // [])[] | select(.app == $agent) | .riskScore // 0] | add / length)
  }))
}'

# Alternative: Get KPIs per agent
for agent_id in $(olakai agents list --json | jq -r '.[].id'); do
  echo "Agent: $agent_id"
  olakai activity kpis --agent-id $agent_id --json | jq '.kpis[] | {name, value}'
done

Report Template

# Agent Comparison Report
Generated: [DATE]
Agents Compared: [COUNT]

## Activity Volume
| Agent | Events | Tokens | Avg Tokens |
|-------|--------|--------|------------|
| [NAME] | [COUNT] | [TOKENS] | [AVG] |

## KPI Comparison
| KPI | [AGENT1] | [AGENT2] | [AGENT3] |
|-----|----------|----------|----------|
| Executions | [VAL] | [VAL] | [VAL] |
| Compliance | [VAL]% | [VAL]% | [VAL]% |
| ROI | $[VAL] | $[VAL] | $[VAL] |

## Risk Profile
[ASCII GROUPED BAR CHART]

## Activity Trend by Agent
[ASCII MULTI-LINE CHART]

Example Output

# Agent Comparison Report
Generated: 2025-01-21
Agents Compared: 4

## Activity Volume
| Agent | Events | Tokens | Avg Tokens |
|-------|--------|--------|------------|
| code-assistant | 245 | 98,450 | 402 |
| data-analyzer | 189 | 156,230 | 827 |
| chat-support | 312 | 78,540 | 252 |
| test-agent | 54 | 12,340 | 229 |

## KPI Comparison
| KPI | code-assist | data-analyze | chat-support |
|-----|-------------|--------------|--------------|
| Compliance | 99.5% | 98.2% | 99.8% |
| Avg Risk | 1.8 | 3.2 | 1.2 |
| Time Saved | 18.5 hrs | 12.3 hrs | 8.7 hrs |

## Risk Profile by Agent
           Low    Medium    High
code-assist ████████████████████ █    │   92%  6%  2%
data-analyze ██████████████████ ████  ██  85% 10%  5%
chat-support █████████████████████ │   │   97%  2%  1%
test-agent  ███████████████████ ██    │   90%  8%  2%

ASCII Visualization Functions

Bar Chart Generator

To create horizontal bar charts, use this pattern:

# Generate bar chart from jq output
cat /tmp/events.json | jq -r '
  [.prompts[].model] | group_by(.) | map({model: .[0], count: length}) |
  sort_by(-.count) |
  (max_by(.count).count) as $max |
  .[] |
  "\(.model | .[0:15] | . + " " * (15 - length))  " +
  ("█" * ((.count / $max * 40) | floor)) +
  " \(.count)"
'

Example output:

gpt-4o           ████████████████████████████████████████ 45
gpt-4o-mini      █████████████████████████ 28
claude-3-5       ██████████████████ 20

Percentage Bar

# Show percentage with visual bar
echo "Compliance: ████████████████████░░░░░ 85%"

Pattern:

[LABEL]: [FILLED █ * percentage/4][EMPTY ░ * (25-filled)] [VALUE]%

Trend Indicators

↑ +7%   (increase)
↓ -3%   (decrease)
→ 0%    (stable)

Quick Reference Commands

# Usage Summary
olakai activity list --limit 100 --json | jq '{
  events: (.prompts | length),
  tokens: ([.prompts[].tokens // 0] | add),
  models: ([.prompts[].model] | unique)
}'

# KPI Snapshot
olakai activity kpis --json | jq '.kpis[] | {name, value, unit}'

# Risk Summary
olakai activity list --limit 100 --json | jq '{
  high_risk: ([.prompts[] | select(.riskScore >= 7)] | length),
  blocked: ([.prompts[] | select(.status == "blocked")] | length)
}'

# Agent List
olakai agents list --json | jq '.[] | {id, name}'

# Per-Agent KPIs
olakai activity kpis --agent-id AGENT_ID --json

# Time-Series Data
olakai activity kpis --period daily --json
olakai activity kpis --period weekly --json

Generating a Complete Report

Follow this workflow for any report type:

# 1. Determine scope
AGENT_ID="your-agent-id"  # or leave empty for all
LIMIT=100

# 2. Collect data
olakai activity list --limit $LIMIT --include-analytics --json > /tmp/activity.json
olakai activity kpis --agent-id $AGENT_ID --json > /tmp/kpis.json
olakai agents list --json > /tmp/agents.json

# 3. Process and format (example for usage summary)
echo "# Usage Summary Report"
echo "Generated: $(date +%Y-%m-%d)"
echo ""
echo "## Overview"
cat /tmp/activity.json | jq -r '"| Metric | Value |
|--------|-------|
| Total Events | \(.prompts | length) |
| Total Tokens | \([.prompts[].tokens // 0] | add) |
| Unique Models | \([.prompts[].model] | unique | length) |"'

Error Handling

No Data Available

# Check if events exist
olakai activity list --limit 1 --json | jq '.prompts | length'

# If 0, inform user:
# "No events found. Ensure your agent is sending events to Olakai."

Agent Not Found

# Verify agent exists
olakai agents list --json | jq '.[] | select(.id == "AGENT_ID")'

# If empty, list available agents:
olakai agents list --json | jq '.[] | {id, name}'

Missing Permissions

# Re-authenticate if needed
olakai logout && olakai login
olakai whoami  # Verify

Best Practices

1. Always use --json flag for programmatic processing
2. Pipe through jq for clean data extraction
3. Cache data locally when generating multi-section reports
4. Include timestamps in all reports
5. Show data freshness - how recent the events are
6. Handle empty states gracefully with informative messages