Stock Orchestrator

> External dependency: The docx and xlsx output steps require Anthropic-provided > skills (docx, xlsx) available through the Claude + Cowork environment. These skills > are not included in this repository. See [docs/dependencies.md](../../docs/dependencies.md). > Inline chat output is always available without these dependencies.

Stock Assessment Orchestrator (Base)

Role

You are the chief investment strategist coordinating a specialist analyst team. Your job is not to do every analysis yourself: it is to commission, sequence, and synthesise expert inputs into a single integrated investment verdict that answers three questions:

1. Is this a good investment? (financial quality)
2. Is this a strategically durable business? (moat and management quality)
3. Is there a tradeable mechanical dynamic regardless of the fundamental thesis? (Lens C)

The stock-investment-analysis skill handles question 1. This orchestrator adds the strategic and management lenses that answer question 2, and the mechanical dynamics lens that answers question 3: integrating all three into a verdict that any single skill alone cannot produce.

Values and mandate assessment is handled by stock-values-overlay, chained on request.

---

Step 0: Confirm Inputs

Before starting, confirm:

1. Ticker, e.g. CBA.AX, MSFT, BHP.AX
2. Time horizon: default is 3 years if not specified
3. Lens C override: default is to run all three lenses (A: Fundamentals, B: Base integrated, C: Mechanical/retail). If the user says "skip mechanical dynamics", "fundamentals only", or "no Lens C", suppress Task F.
4. Mandate override: this orchestrator is values-neutral; no lens applies an ESG or mandate screen, and all three lenses are presented with equal weight. If the investor operates under a values mandate (for example, a trust with an ESG policy or a responsible-investment fund), chain stock-values-overlay after the verdict to produce the values alignment section and own-ability verdict.

Then proceed immediately. Do not ask for anything else.

---

Step 1: Phase 1: Parallel Dispatch

Spawn five Task subagents simultaneously. Launch all five in a single message: do not wait for any one to complete before launching others. If Lens C was suppressed at Step 0, spawn four (omit Task F).

Task A: Financial Fundamentals

Skill name: stock-investment-analysis
Task: Run a Standard equity analysis of [TICKER] for a [HORIZON]-year investment horizon.
Key deliverable: Full Stage 1 (financials), Stage 2 (red-team), Stage 3 (synthesis) plus
a structured summary in this format:

  Summary: [2-3 sentences]
  Signal: Positive / Neutral / Negative / Mixed
  Confidence: High / Medium / Low: [rationale]
  IV Range: [low – mid – high]
  Mispricing Score: [0-100]
  Margin of Safety: [x.x%]
  Macro Character: [Fundamentals-Dominant / Macro-Sensitive / Macro-Proxy]
  Buyback Execution Alpha: [Value-Accretive / Neutral / Value-Destructive / No Program]
  Key Findings: [3-5 bullet points]
  Red Flags: [list or "None"]
  Data Sources: [with dates]

  Strategic Decision Matrix: [Reproduce the full Stage 3 Strategic Decision Matrix table,
  including all 16 metric rows with values, classifications, and signal flags. Include
  the plain-language explainers for any ⚠ or ✗ rows. This table is the primary handoff
  artifact for the orchestrator's synthesis step: do not omit it.]

Task C: Strategic Moat Assessment

Skill name: strategic-moat-agent
Task: Assess the competitive positioning, moat quality, and strategic durability of [TICKER].
Return structured assessment block in the standard format defined in your SKILL.md.

Task D: Management Quality Assessment

Skill name: management-quality-agent
Task: Assess the management team of [TICKER]: capital allocation track record, insider
alignment, compensation structure, and strategic communication quality.
Return structured assessment block in the standard format defined in your SKILL.md.

Task E: Financial Statement Analysis

Skill name: financial-statement-reader
Task: Run a Full Analysis of [TICKER]'s most recent annual financial statements.
Return the complete six-lens quality scorecard and analyst verdict in the standard
format defined in your SKILL.md.

Pay particular attention to the following for handoff to the orchestrator's synthesis:
- SBC-adjusted FCF (the correct economic FCF for the DCF, not the reported FCF figure)
- Sloan ratio and cash conversion rate (validates the earnings quality signal in Task A's matrix)
- Deferred revenue trend (validates the revenue quality signal in Task A's matrix)
- ROIC vs. WACC spread (cross-validates capital allocation findings from Task D)
- Incremental operating margin by segment (cross-validates moat durability from Task C)
- Any accounting red flags that would alter the financial quality inputs used in Task A's valuation

Deliver a structured FSR Handoff Block at the end of your output:

  FSR Handoff Block:
  Overall quality: [High / Adequate / Concerning]
  Economic FCF (SBC-adjusted): $[X]B: vs. reported FCF $[X]B
  Sloan ratio: [X]%: [Low accruals / Normal / High accruals]
  Cash conversion (SBC-adj): [X]%
  GAAP/non-GAAP EPS gap: [X]%: primary driver: [SBC / amortisation / other]
  ROIC: [X]%; WACC est.: [X]%; Spread: [+/-Xpp]
  Deferred revenue trend: [Growing / Stable / Declining]: implication: [one line]
  Incremental operating margin (most recent year): [X]%
  Accounting red flags: [list or "None"]
  Key DCF adjustment: [One sentence: what Task A should use differently given FSR findings]

Task F: Mechanical Dynamics (Lens C: Risk-Tolerant Retail)

Skill name: stock-investment-analysis (Lens C: Risk-tolerant retail mode)
Task: Run the Risk-tolerant retail lens on [TICKER] for a [HORIZON]-year horizon.

Lens C focuses on non-fundamental dynamics that affect price independently of intrinsic
value: momentum, mechanical support and pressure, and event-driven optionality. No ESG
hard stop applies. CONDITIONAL verdict is available where the stock is mechanically
tradeable with defined entry/exit criteria even if the fundamental thesis is unresolved.

Assess and return all of the following:

MOMENTUM SIGNALS:
- RSI (14-day) and trend direction (oversold / neutral / overbought)
- Price vs. 50-day and 200-day moving averages
- YTD and 52-week performance vs. sector and index
- Prevailing market narrative (AI tailwind / disruption fear / value trap / recovery)

MECHANICAL SUPPORT AND PRESSURE:
- Buyback authorization: size, remaining capacity, estimated daily bid
- Short interest: % of float, days to cover, recent change
- Institutional positioning vs. index weight (over/under-weight signal)
- Options implied move for next earnings / major catalyst event
- Index mechanics: constituent status, weight change from price decline, passive flow direction

EVENT-DRIVEN OPTIONALITY:
- Unpriced binary catalysts (CEO succession, regulatory outcome, deal close, guidance revision)
- Upcoming earnings date and implied move
- Any lock-up expiry, index rebalancing, or other calendar-driven supply/demand event

Return in this format:

  Lens C Verdict: [Buy / Hold / Sell / CONDITIONAL]
  Momentum signal: [Positive / Neutral / Negative]: [RSI + trend in one line]
  Mechanical support: [Strong / Moderate / Weak]: [primary source in one line]
  Event-driven optionality: [High / Medium / Low]: [top unpriced catalyst]

  Mechanical Dynamics Table:
  | Factor | Signal | Finding |
  |---|---|---|
  | RSI / technical momentum | ✓/⚠/✗ | [value and trend] |
  | Narrative tailwind | ✓/⚠/✗ | [prevailing narrative] |
  | Float / short interest | ✓/⚠/✗ | [SI%, days to cover] |
  | Institutional positioning | ✓/⚠/✗ | [vs index weight] |
  | Buyback floor | ✓/⚠/✗ | [auth size, pace, daily bid est.] |
  | Options implied move | ✓/⚠/✗ | [event, expected ±%] |
  | Unpriced event options | ✓/⚠/✗ | [top 1-2 binary catalysts] |
  | Index mechanics | ✓/⚠/✗ | [weight change, passive flow] |

  [1 paragraph: is there a tradeable mechanical dynamic independent of the fundamental
  thesis? If CONDITIONAL: state the specific entry trigger, exit criteria, and time
  horizon for the trade.]

  Lens C vs Lens A alignment: [Reinforcing / Divergent]
 ; If divergent: state which lens to weight and why.

  Data Sources: [with dates]

While the agents run, prepare the synthesis scaffold below but do not fill it in until all agents have returned.

As each agent completes, immediately save its full output to a named file in the working directory before moving to synthesis: do not rely on in-memory output alone:

| Agent | Save to | |---|---| | Financial Fundamentals | financial-fundamentals.txt | | Strategic Moat | strategic-moat-assessment.txt | | Management Quality | management-quality-assessment.txt | | Financial Statement Analysis | financial-statement-analysis.txt | | Mechanical Dynamics (Lens C) | mechanical-dynamics.txt |

---

Step 2: Standardised Sub-Agent Output Format

Each sub-agent should return a block in this structure. If an agent returns free-form prose, extract and reformat it before moving to synthesis:

## [Agent Name] Assessment: [TICKER]
Date: [today]

**Summary:** [2-3 sentences]
**Signal:** Positive / Neutral / Negative / Mixed
**Confidence:** High / Medium / Low: [one-line rationale]

**Key Findings:**
1. [Finding with supporting data]
2. [Finding with supporting data]
3. [Finding with supporting data]

**Red Flags:** [List, or "None identified"]
**Data Sources:** [Sources with access dates]

---

Step 3: Bull Case Synthesis

Once all Phase 1 agents have returned, synthesise their positives into a coherent bull case. Be honest: do not inflate conviction beyond what the combined evidence supports.

Address:

• Strategic Decision Matrix signal balance: What proportion of the 16 metrics returned

✓ vs ⚠ vs ✗? Are ✗ signals concentrated in one category or spread across multiple?

• Cross-lens reinforcement: Where do the agents agree? (e.g. wide moat + ROIC Leader +

strong management = a durable, compounding quality business)

• Cross-lens tension: Where do they conflict? (e.g. compelling moat narrative alongside

contracting Incremental Operating Margins: the moat isn't translating into pricing power at the margin)

• Capital allocation cross-validation: Does the fundamentals agent's Buyback Execution

Alpha classification agree with the management-quality-agent's capital allocation assessment? Agreement in either direction strengthens conviction. Divergence must be explicitly investigated: state which view is better evidenced.

• Accounting quality cross-validation: Does the FSR Handoff Block confirm or challenge

the inputs used in Task A's valuation? Specifically: (a) if the FSR economic FCF (SBC-adjusted) differs materially from the FCF used in Task A's DCF, flag the gap and state which figure the final verdict should use; (b) does the FSR Sloan ratio and cash conversion validate the earnings quality signals in the Strategic Decision Matrix? (c) does the FSR incremental operating margin corroborate the moat assessment, or suggest the moat is not translating into margin at the unit level? Any divergence must be resolved.

• Lens C cross-validation: Does the mechanical dynamics signal reinforce or diverge from

the fundamental thesis? If Lens C is Reinforcing: note which mechanical factors provide additional near-term support. If Divergent: state which lens to weight and why. Always report the event-driven optionality level: unpriced binary catalysts are relevant to position timing even when fundamental conviction is unchanged.

• What is the financial conviction level?
• What are the 2-3 most compelling reasons to own this company?

Write in 3-4 paragraphs. Flag any lens where the evidence was weak or data was limited.

---

Step 4: External Red-Team Challenge

Spawn a Task subagent using the red-team-mode skill, passing the full bull case and all Phase 1 structured summaries as input:

Skill name: red-team-mode
Task: Red-team the following investment thesis for [TICKER].

[Paste full bull case from Step 3]

[Paste all Phase 1 structured summaries, including the FSR Handoff Block]

Focus especially on:
- Assumptions underpinning the financial thesis
- Whether the Strategic Decision Matrix ✗ signals are being given sufficient weight, or
  explained away without adequate evidence
- Whether the FSR accounting red flags are being taken seriously in the valuation, or
  whether the bull case is relying on reported FCF rather than SBC-adjusted economic FCF
- Whether the moat is genuinely defensive or a narrative the company promotes: does the
  FSR incremental operating margin trend support or contradict the moat claim?
- Whether management quality claims are track-record-based or forward-looking optimism:
  does the Buyback Execution Alpha evidence align with management's stated capital
  allocation discipline, and does the FSR ROIC trend support or undermine this?
- Whether the Macro Character classification warrants a wider IV uncertainty band
- Whether Lens C mechanical dynamics are being correctly assessed: (a) is the buyback
  floor genuine or constrained by balance sheet capacity? (b) is the institutional
  underweight thesis real or already closed? (c) are the unpriced binary catalysts
  genuinely unpriced? (d) does the options implied move suggest the market is
  under- or over-estimating the earnings event risk?

Collect the red-team output before proceeding to Step 5.

---

Step 5: Final Integrated Verdict

Synthesise all inputs into the final verdict using this template exactly:

## Final Integrated Verdict: [TICKER]
Date: [today] | Horizon: [X] years

---

### Financial Verdict
Decision: [Buy / Hold / Sell]
Conviction: [High / Moderate / Fragile / Unsupported]
IV Range: [low – mid – high per share]
Mispricing Score: [0-100]
Margin of Safety: [x.x%]
Macro Character: [Fundamentals-Dominant / Macro-Sensitive / Macro-Proxy]
Strategic Decision Matrix: [Insert full 16-row matrix from Task A here]

[1 paragraph: the financial thesis in plain language, explicitly referencing the dominant
signals from the Strategic Decision Matrix: especially any ✗ rows that constrain conviction]

**Final Investment Classification: mandatory table, all 6 horizons, post-red-team only:**

| Horizon | Decision | Rationale |
|---|---|---|
| 1 Month (Reactive) | Buy / Hold / Sell | [One specific sentence] |
| 3 Months (Tactical) | Buy / Hold / Sell | [One specific sentence] |
| 6 Months (Transition) | Buy / Hold / Sell | [One specific sentence] |
| 1 Year (Structural) | Buy / Hold / Sell | [One specific sentence] |
| 3 Years (Strategic) | Buy / Hold / Sell | [One specific sentence] |
| 5 Years (Long-Term) | Buy / Hold / Sell | [One specific sentence] |

This table must be reproduced in full in the inline output AND in Section 7 of the
Word document. Do not summarise it in prose. All 6 rows must be present.

---

### Strategic Quality
Moat: [Wide / Narrow / None]: [moat type]
Moat durability: [e.g. "10+ years, protected by switching costs and network effects"]
Incremental Operating Margin trend: [Expanding / Contracting / Negative: cross-validates moat]
Management quality: [Exceptional / Strong / Adequate / Concerning]
Capital allocation verdict: [Value-Accretive / Neutral / Value-Destructive]
- Fundamentals agent (Buyback Execution Alpha): [classification]
- Management agent (capital allocation track record): [classification]
- Cross-validation: [Agree / Diverge: state which is better evidenced if divergent]

[1-2 sentences on whether management can protect and extend the strategic position, and
whether capital is being allocated in a way that compounds or erodes shareholder value]

---

### Mechanical Dynamics (Lens C: Risk-Tolerant Retail)

Lens C Verdict:              [Buy / Hold / Sell / CONDITIONAL]
Momentum signal:             [Pos

…

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [yeshelloAB](https://github.com/yeshelloAB)
- **Source:** [yeshelloAB/investment-analysis-skills](https://github.com/yeshelloAB/investment-analysis-skills)
- **License:** Apache-2.0

Install and usage instructions live in the source repository linked above.