Earnings Preview Agent Skill

Preflight

Run the canonical pre-flight sequence — MCP health probe, ticker resolution, workspace

style.md override, memory load, and coverage check. See contracts/preflight.md.

Office dependency probe (FR-043) — this skill produces .pptx via Bash + python-pptx:

• Live Office session? If mcp__office__* tools are present (Cowork), drive the live

document instead of headless Python.

• Python library: Bash: python3 -c "import pptx" — if exit ≠ 0, fall back to .md

slide spec with data_availability: degraded + python_pptx_missing: true.

• LibreOffice: Bash: which soffice for structural validation and PDF export.

If the Python library is absent, report the exact remediation:

pip install python-pptx and produce the .md degraded fallback per

contracts/office-tooling.md.

Include the X-Agentii-Trace header on every tool call per

contracts/x-agentii-trace-header.md.

Triggers

• generate earnings preview deck
• build earnings preview presentation
• create quarterly earnings slides
• earnings preview pptx
• earnings summary presentation
• consensus estimates presentation
• earnings surprise summary deck
• quarterly results presentation
• earnings catalyst calendar slides
• pre-earnings analyst deck

Defaults

| Parameter | Default | Notes |

|-----------|---------|-------|

| slide_count | 4-6 | Title, Company Overview, Consensus Estimates, Historical Surprises, Catalysts, Outlook |

| lookback_quarters | 4 | Trailing 4 quarters for trend analysis |

| peer_count | 3-5 | From search_companies sector peers |

| source_footers | required | Every slide has standard agentii citation footer |

| template | institutional-default | Dark header bar, agentii blue accent, 12pt body |

Methodology

Retrieval Scope

This skill performs structured data retrieval (earnings calendar, XBRL facts, company profile) with simple lookups — no unstructured document search. retrieval_scope: structured_only applies. See references/formula-sheet.md for presentation structure guidelines.

Retrieval Strategy

See contracts/retrieval.md for the canonical decision tree; skill-specific retrieval detail is in references/methodology.md.

Temporal Scope

Default: 4 fiscal quarters (max 8). Trailing 4 quarters captures current estimates and YoY comparisons. Maximum 8 quarters for analysts who want 2-year trend context on the estimates slide.

Tool Allowlist

See frontmatter allowed_tools. This skill produces a polished .md slide-deck specification; .pptx rendering is available via the companion financial-analysis:pptx-author skill (separate install; see contracts/office-tooling.md).

Protocol

Step-by-step execution detail is in references/methodology.md.

Deliverable Chain

Inputs → Build → Validate → Output → Next

• Inputs: resolved ticker + earnings calendar, consensus estimates, and trailing XBRL facts (search_earnings_calendar, search_xbrl_facts, search_companies, get_company_profile).
• Build: write a self-contained Python script using python-pptx that creates the 4–6 slide .pptx deck per ## Output Structure. Execute via Bash: python3 script.py. Verify the .pptx file exists. If python-pptx is absent, fall back to .md slide spec per contracts/office-tooling.md.
• Validate: run the ## Validation Gates below.
• Output: write the artifact path per ## Output File.
• Next: append to agentii.md; hand off to a downstream pitch/review skill if requested.

Validation Gates

• slide count: between 4 and 6. *If failed*: If outside range: refuse delivery.
• estimates slide: includes consensus, high, and low estimates. *If failed*: If missing: flag in Coverage Gaps.
• source footers: every slide has source_footer with standard agentii citation. *If failed*: If any missing: refuse delivery.
• peer comparison: has >= 3 peers. *If failed*: If < 3: flag in Coverage Gaps.

Tool Fallbacks

Per-tool failure modes and fallback actions are tabulated in references/tool-fallbacks.md.

Output File

Primary deliverable: {ticker}/{YYYY-MM-DD_HHMM}_earnings-preview_{affix}.pptx — real PowerPoint binary via Bash + python-pptx per contracts/office-tooling.md. Degraded fallback: {ticker}/{YYYY-MM-DD_HHMM}_earnings-preview_{affix}.md when python-pptx is absent (FR-044).

Output Structure

The deliverable is a structured markdown report written to the path in ## Output File. Full section-by-section template (headings, tables, and field definitions) lives in references/output-structure.md. Required elements:

• Executive Summary — headline conclusions (≤200 words).
• Core analysis sections — per this skill's methodology and analyst modes.
• Data classification — tag findings [FACT] / [DEDUCTED] / [VIEW] per contracts/snapshot-synthesis.md.
• Coverage Gaps & Citations — inline /v/ citations are PRIMARY (immediately after each fact); the bottom Citations section is a non-duplicative roll-up index.
• Output frontmatter — emit the FR-090 structured block per contracts/output-frontmatter-schema.md.

Citations & memory: follow contracts/citation-and-memory.md — ≥1 citation per 200 words; every material fact, table row, and metric is immediately followed by its inline clickable https://agentii.ai/v/{ticker}/{citation_id}/{N} link; a bottom Citations section provides a non-duplicative roll-up index; the closing TUI reply includes a compact Key Citations list (headline 5–10 facts) of clickable /v/ URLs; and append the run to agentii.md per contracts/agentii-md-schema.md.

Memory & Snapshot

• Memory load (pre-flight): load prior workspace context for the ticker before retrieval — see contracts/memory-load.md.
• Structured output frontmatter: emit the FR-090 block (key_metrics, conclusions, facts_count, deducted_count, views_count, citation_count) per contracts/output-frontmatter-schema.md.
• Snapshot synthesis: after writing the deliverable, update the two-tier snapshot and classify findings as [FACT]/[DEDUCTED]/[VIEW] — see contracts/snapshot-synthesis.md.
• Session archival: record the run under sessions/{YYYY-MM-DD}/ and update sessions/INDEX.md per contracts/session-format.md.

Final Summary (TUI)

End the closing chat reply with a compact Key Citations list (headline 5–10 facts), each a clickable https://agentii.ai/v/{ticker}/{citation_id}/{N} link, so the user can cmd+click straight to the exact SEC page. See contracts/citation-and-memory.md.

Error Handling

| Failure Mode | Detection | Action | User-Facing Message |

|---|---|---|---|

| Missing earnings data | search_earnings_calendar returns empty | Use search_xbrl_facts for historical actuals only; flag estimates as unavailable | "Consensus estimates not available for {ticker}; presentation based on historical actuals only." |

| Partial data | <80% expected fields returned | Proceed with coverage gaps section | "Presentation based on partial data; see Coverage Gaps." |

| Sector mismatch | Peer sector != target sector | Filter out mismatched peers | "Removed {n} peer(s) due to sector mismatch." |

| Insufficient history | <4 quarters of data available | Downgrade to limited-history presentation (3 slides min) | "Limited historical data available; presentation adjusted." |

| MCP unreachable | agentii Preflight probe fails | Halt with actionable error | "agentii data plane unreachable; check connection and AGENTII_API_KEY." |

| Office backend unreachable | All 3 office backends fail Preflight | Halt with AGENTII_OFFICE_UNREACHABLE | "No office backend available. Options: (a) set AGENTII_API_KEY for agentii-office, (b) pip install python-pptx, (c) install OfficeCLI." |

| Knowledge Store unavailable | get_entity_knowledge returns 503 | Fall back to get_company_profile + search_companies; flag with knowledge_store_degraded: true | "Knowledge Store not yet available; peer analysis based on filing-derived entity context." |