report — Claude Code command: reports and installation
Generate task, sprint, and performance reports with /task-report. Use it for a standup, project review, or a closer look at blockers and dependencies.
npx claude-code-templates@latest --command orchestration/reportWhat report does
Run /task-report with a report type to create an executive summary, sprint report, standup, or analysis of risks, dependencies, and agent workload. Options let you choose a project, time period, and output format such as Markdown, HTML, or PDF.
The command reads data from TASK-STATUS-TRACKER.yaml files and includes completed tasks in historical metrics. Time calculations use business hours by default. Charts need Unicode support in the terminal, so their display depends on the terminal environment.
To install it in Claude Code, run npx claude-code-templates@latest --command orchestration/report. The command file is .claude/commands/report.md.
What it helps with
- summarize task progress and blockers
- compare sprint metrics across periods
- trace dependencies and identify bottlenecks
- review agent workload and performance
- format a status report for a chosen output
Example prompts for the agent
- Create a standup covering completed, ongoing, and blocked work:
/task-report standup. - Prepare a sprint report with velocity and burndown metrics:
/task-report sprint --include-velocity --include-burndown. - Show the project dependencies as a graph:
/task-report dependencies --visual.
How to install report
1claude-code-templates
npx claude-code-templates@latest --command orchestration/report2Manually: put the files in place
.claude/commands/report.mdCall it with /report. For all projects use ~/.claude/commands/
Manually: put the files in place
~/.codex/prompts/report.mdCall it with /prompts:report. Codex marks this format deprecated in favour of skills
Manually: put the files in place
.github/prompts/report.prompt.mdManually: put the files in place
.gemini/commands/report.tomlGemini stores commands as TOML: the prompt text goes into the prompt field
Questions about report
How do I install report in Claude Code?
Run npx claude-code-templates@latest --command orchestration/report. The command file is .claude/commands/report.md.
What does the report command do?
It creates reports on task execution, sprints, performance, risks, and dependencies. You can select a project, period, and output format.
What data does report use, and what do its charts require?
Reports draw on TASK-STATUS-TRACKER.yaml files. Charts require Unicode support in the terminal.
Source text
report.md · MIT
--- description: "Generate comprehensive reports on task execution, progress, and metrics." --- # Task Report Command Generate comprehensive reports on task execution, progress, and metrics. ## Usage ``` /task-report [report-type] [options] ``` ## Description Creates detailed reports for project management, sprint reviews, and performance analysis. Supports multiple report types and output formats. ## Report Types ### Executive Summary ``` /task-report executive ``` High-level overview for stakeholders with key metrics and progress. ### Sprint Report ``` /task-report sprint --date 03_15_2024 ``` Detailed sprint progress with burndown charts and velocity. ### Daily Standup ``` /task-report standup ``` What was completed, in progress, and blocked. ### Performance Report ``` /task-report performance --period week ``` Team and individual performance metrics. ### Dependency Report ``` /task-report dependencies ``` Visual dependency graph and bottleneck analysis. ## Output Examples ### Executive Summary Report ``` EXECUTIVE SUMMARY - Authentication System Project ================================================ Report Date: 2024-03-15 Project Start: 2024-03-13 Duration: 3 days (60% complete) KEY METRICS ----------- • Total Tasks: 24 • Completed: 12 (50%) • In Progress: 3 (12.5%) • Blocked: 2 (8.3%) • Remaining: 7 (29.2%) TIMELINE -------- • Original Estimate: 5 days • Current Projection: 5.5 days • Risk Level: Low HIGHLIGHTS ---------- ✓ Core authentication API completed ✓ Database schema migrated ✓ Unit tests passing (98% coverage) BLOCKERS -------- ⚠ Payment integration waiting on external API ⚠ UI components need design approval NEXT MILESTONES -------------- → Complete JWT implementation (Today) → Integration testing (Tomorrow) → Security audit (Day 4) ``` ### Sprint Burndown Report ``` /task-report burndown --sprint current ``` ``` SPRINT BURNDOWN - Sprint 24 =========================== Tasks Remaining by Day: Day 1: ████████████████████ 24 Day 2: ████████████████ 20 Day 3: ████████████ 15 (TODAY) Day 4: ████████ 10 (projected) Day 5: ████ 5 (projected) Velocity Metrics: - Average: 4.5 tasks/day - Yesterday: 5 tasks - Today: 3 tasks (in progress) Risk Assessment: ON TRACK ``` ### Performance Report ``` TEAM PERFORMANCE REPORT - Week 11 ================================= By Agent: ┌─────────────────┬────────┬───────────┬─────────┬────────────┐ │ Agent │ Completed │ Avg Time │ Quality │ Efficiency │ ├─────────────────┼────────┼───────────┼─────────┼────────────┤ │ dev-frontend │ 8 │ 3.2h │ 95% │ 125% │ │ dev-backend │ 6 │ 4.1h │ 98% │ 110% │ │ test-developer │ 4 │ 2.8h │ 100% │ 115% │ └─────────────────┴────────┴───────────┴─────────┴────────────┘ By Task Type: - Features: 12 completed (avg 3.8h) - Bugfixes: 4 completed (avg 1.5h) - Tests: 8 completed (avg 2.2h) Quality Metrics: - First-time pass rate: 88% - Rework required: 2 tasks - Blocked time: 4.5 hours total ``` ## Customization Options ### Time Period ``` /task-report summary --from 2024-03-01 --to 2024-03-15 /task-report summary --last 7d /task-report summary --this-month ``` ### Specific Project ``` /task-report sprint --project authentication_system ``` ### Format Options ``` /task-report executive --format markdown /task-report executive --format html /task-report executive --format pdf ``` ### Include/Exclude ``` /task-report summary --include completed,qa /task-report summary --exclude on_hold ``` ## Specialized Reports ### Critical Path Analysis ``` /task-report critical-path ``` Shows tasks that directly impact completion time. ### Bottleneck Analysis ``` /task-report bottlenecks ``` Identifies tasks causing delays. ### Resource Utilization ``` /task-report resources ``` Shows agent allocation and availability. ### Risk Assessment ``` /task-report risks ``` Identifies potential delays and issues. ## Visualization Options ### Gantt Chart ``` /task-report gantt --weeks 2 ``` ### Dependency Graph ``` /task-report dependencies --visual ``` ### Status Flow ``` /task-report flow --animated ``` ## Automated Reports ### Schedule Reports ``` /task-report schedule daily-standup --at "9am" /task-report schedule weekly-summary --every friday ``` ### Email Reports ``` /task-report executive --email [email protected] ``` ## Comparison Reports ### Sprint Comparison ``` /task-report compare --sprint 23 24 ``` ### Week over Week ``` /task-report trends --weeks 4 ``` ## Examples ### Example 1: Morning Status ``` /task-report standup --format slack ``` Generates Slack-formatted standup report. ### Example 2: Sprint Review ``` /task-report sprint --include-velocity --include-burndown ``` Comprehensive sprint metrics for review meeting. ### Example 3: Blocker Focus ``` /task-report blockers --show-dependencies --show-resolution ``` Deep dive into what's blocking progress. ## Integration Features ### Export to Tools ``` /task-report export-jira /task-report export-asana /task-report export-github ``` ### API Endpoints ``` /task-report api --generate-endpoint ``` Creates API endpoint for external access. ## Best Practices 1. **Daily Reviews**: Run standup report each morning 2. **Weekly Summaries**: Generate performance reports on Fridays 3. **Sprint Planning**: Use velocity trends for estimation 4. **Stakeholder Updates**: Schedule automated executive summaries ## Report Components Each report can include: - Summary statistics - Timeline visualization - Task lists by status - Agent performance - Dependency analysis - Risk assessment - Recommendations - Historical trends ## Notes - Reports use data from all TASK-STATUS-TRACKER.yaml files - Completed tasks are included in historical metrics - Time calculations use business hours by default - All times shown in local timezone - Charts require terminal unicode support
© davila7 · MIT · davila7/claude-code-templates