company-research-workflow

Company Research Workflow

Use this skill when the user asks for a complete company-level investment research run. It composes narrower skills and keeps the persistent archive as the shared memory layer.

Workflow

1. Run research-target-resolver.
2. Run research-evidence-archive to create or load the target archive.
3. Decide which collection skills are needed for the user's question.
4. Run independent collection skills in parallel when possible: company-news-research, company-filing-research, earnings-call-analysis, financial-snapshot-analysis, market-signal-analysis, competitive-landscape-analysis, investment-risk-diligence, and valuation-scenario-analysis when valuation, margin of safety, or price-implied expectations matter.
5. Save all source notes and raw files through research-evidence-archive.
6. Run investment-thesis-synthesis only after evidence coverage is visible.
7. Run portfolio-exposure-review when the user asks about sizing, adding, trimming, holding, watchlists, or portfolio fit.
8. Return the final artifact paths, major findings, gaps, and suggested next research runs.

Read references/intel-example-workflow.md for the default Intel-like scenario.

Skill Selection

• User asks "what happened recently": prioritize target, archive, news, filings, earnings, market, then thesis.
• User asks "is the business improving": prioritize filings, earnings, financials, competitors, risks, then thesis.
• User asks "why did the stock move": prioritize news, earnings, market, then filings if the event is disclosure-driven.
• User asks "how does it compare": prioritize target, financials, competitors, market, then thesis.
• User asks "what is it worth" or "is it priced in": prioritize target, financials, filings, peers, market, risks, valuation scenarios, then thesis.
• User asks "what can go wrong": prioritize target, filings, earnings, news, financials, market, risk diligence, then thesis.
• User asks "how much should I own" or "does this fit my portfolio": prioritize target, thesis, valuation, risks, market, portfolio context, then portfolio exposure review.

Output

Return:

• skills executed
• archive paths touched
• evidence coverage table
• risk diligence path when created
• valuation scenario path when created
• final memo or decision-frame path
• portfolio review path when created
• unresolved gaps
• recommended follow-up skills

Quality Gate

Before finishing:

• do not synthesize before evidence coverage is known
• store reusable materials in research/ instead of leaving them only in chat
• label current-data retrieval failures
• keep user-facing conclusions tied to archived evidence
• avoid autonomous trading, broker actions, or certainty about future returns