| name | prepare-earnings |
| description | Prepare a source-linked pre-earnings brief or post-earnings thesis update using Diligence Stack research and separately entitled equity research when available. Use for earnings previews, KPI expectations, guidance bridges, management questions, estimate changes, transcript analysis, or determining what an earnings release changes. |
Prepare Earnings
Use the MCP for research context and thesis history; use current primary company materials for same-day reported results.
Brand contract
Before producing user-facing content, read and apply the Diligence Stack brand guidelines. Use its color, typography, logo, citation, and link defaults unless the user explicitly requests different visual styling; its attribution and canonical-link rules always apply.
Choose the mode
- Pre-earnings: identify the debate, critical KPIs, expectations, scenarios, questions, and thesis-moving thresholds.
- Post-earnings: compare reported results and guidance with prior expectations, explain the operational drivers, and update the thesis and model.
Workflow
- Confirm company, ticker, quarter, reporting date, mode, and the user's decision context.
- Follow
diligence-research. Retrieve recent Diligence Stack company/vendor analysis, market context, watch items, and falsification conditions.
- If
equity-research is authorized, sample dated views from multiple publishers. Keep consensus, company guidance, and Diligence Stack assumptions separate.
- For pre-earnings, define the KPI tree and what constitutes upside, in-line, and downside - not just headline revenue/EPS.
- For post-earnings, verify actual results, guidance, and transcript language against a current primary company source. Do not assume the MCP corpus has indexed same-day materials.
- Bridge changes by volume, price, mix, utilization, timing, margin, cost, capacity, or customer behavior.
- Update the model and thesis only where evidence changes an assumption. Flag one-time items and definition changes.
- Produce questions that target remaining uncertainty rather than facts already disclosed.
Use the earnings templates.
Guardrails
Label consensus and estimates with source and date. Price targets and ratings are source views, not facts or recommendations. Mark private-source content and do not expose authenticated links to unauthorized readers.
