Palantir 2026 財報分析：端到端產品測試流程

You are running a source-grounded MadCowork Office E2E workflow.
Communicate final user-facing conclusions in Traditional Chinese, but all tool planning and tool arguments must be concise English.

Topic: Palantir 2026 financial report analysis and outlook (Palantir 2026財報分析及預期).

Known source targets to fetch first:
- Official Q1 2026 release: https://www.businesswire.com/news/home/20260503338048/en/Palantir-Reports-Q1-2026-U.S.-Revenue-Growth-of-104-YY-and-Revenue-Growth-of-85-YY-Raises-FY-2026-Revenue-Guidance-to-71-YY-Growth-and-U.S.-Comm-Revenue-Guidance-to-120-YY-Crushing-Consensus-Expectations
- Official Q1 2026 10-Q: https://investors.palantir.com/files/2026%20Q1%20PLTR%2010-Q.pdf
- Analyst forecast reference: https://stockanalysis.com/stocks/pltr/forecast/

Hard requirements:
1. Do not rely on model memory. Use web_search, web_fetch, and web_crawl as needed before making factual or numerical claims.
2. Gather evidence from multiple sources, including Palantir investor relations / SEC filings and at least one analyst or market forecast source.
3. Create CSV files in this session workspace:
   - pltr_sources.csv with columns: source_type,title,url,published_date,used_for.
   - pltr_financials.csv with columns: company,metric,period,value,unit,source_url,note.
   - pltr_analyst_forecasts.csv with columns: provider,metric,period,forecast,unit,source_url,note.
   - pltr_chart_inputs.csv with columns: chart_name,series,period,value,unit,source_url,note.
   - session_metadata.csv with columns: artifact,type,path,created_at,source_inputs.
4. Every numeric claim used in the deck must have a source_url in one of the CSV files.
5. Include these real numerical analyses if the sources support them:
   - Q1 2026 revenue, Q1 2025 revenue, YoY revenue growth.
   - Q1 2026 gross profit, operating income, net income attributable to common stockholders, adjusted operating income, adjusted free cash flow.
   - FY2026 company revenue guidance range and midpoint.
   - Q2 2026 revenue guidance range.
   - U.S. commercial revenue and U.S. government revenue for Q1 2026.
   - Profitability share analysis: gross profit / revenue, GAAP operating income / revenue, net income attributable to common stockholders / revenue, adjusted operating income / revenue, adjusted FCF / revenue.
   - Analyst forecast comparison: company FY2026 guidance midpoint versus the analyst/market forecast source.
6. Create chart-ready rows in pltr_chart_inputs.csv for:
   - Revenue trajectory: Q1 2025 actual, Q1 2026 actual, Q2 2026 guidance midpoint, FY2026 guidance midpoint, and analyst FY2026 forecast if available.
   - Profitability share: gross margin, operating margin, net income margin, adjusted operating margin, adjusted FCF margin.
   - Segment mix: U.S. commercial revenue versus U.S. government revenue.
7. Call source_validate on pltr_sources.csv with min_official_sources=2 and min_analyst_sources=1. If it fails, record the data gap in the deck instead of inventing analyst facts.
8. Call csv_validate on every CSV before using it for charts or xdeck.
9. Generate a self-contained web presentation through xdeck_generate. When calling xdeck_generate, set out_dir exactly to "pltr-2026-web-deck" so the final file is pltr-2026-web-deck/deck.html. Pass facts_paths, sources_path, chart_data_path, and validation_report_path when available.
10. The deck should be 8 slides, Traditional Chinese, investor / software industry strategy briefing style, with speaker_note on each slide.
11. Include insight about software industry expectations: operational AI budgets, commercial AI adoption, government software demand, and valuation-risk caveat. Ground all numerical claims in CSV rows.
12. Run xdeck_audit and xdeck_screenshots. If the audit finds layout issues, run one xdeck_revise pass.
13. Final answer: list generated artifact paths and briefly explain the evidence chain.

Important: if a source cannot be fetched, record the failed URL in pltr_sources.csv with used_for="attempted but unavailable" and continue with fetchable sources.