# Portfolio Analysis > Analyze listed stocks or an equity portfolio, including CSV holdings, and discover or monitor index-aware mid-cap and catalyst-growth opportunities using current primary-source evidence, normalized per-share fundamentals, valuation, promoter transactions, scenario-based return decomposition, and portfolio risk. Use to review stocks, compare holdings, find opportunities beyond an existing index core, maintain entry/scale/trim/exit watchlists, or produce evidence-based portfolio decisions rather than react to recent ticker returns. - Skill: `ayushgoel/portfolio-analysis` (Agent Skill, multi-file: 7 files) - Install (CLI): `npx skillmds@latest add ayushgoel/portfolio-analysis` - Raw SKILL.md: https://api.skillmd.com/api/skills/ayushgoel/portfolio-analysis/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Marketing & Growth - Author: ayushgoel (https://skillmd.com/u/ayushgoel) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/ayushgoel/portfolio-analysis --- # Portfolio Analysis Analyze prospective returns through business performance, valuation, cash returns, and portfolio fit. Do not infer quality from recent price movement or make a decision from a single ratio. ## Workflow 1. Establish the as-of date, market, horizon, currency, return hurdle, risk needs, liquidity needs, tax constraints, desired actions, and portfolio boundary. Include index funds and externally held equity when the user discloses them. If context is missing, state reasonable illustrative assumptions instead of inventing facts. Do not recommend direct mega-cap exposure as diversification when the user's index funds already provide it. 2. For a portfolio CSV, extract the security identifier, quantity, and acquisition price. Remove zero-quantity rows and combine duplicates using quantity-weighted acquisition cost. Write `.cleaned.csv` beside the input unless that directory is this public skill repository; in that case, use a private temporary workspace or a user-approved output directory. 3. Read [references/two-engine-framework.md](references/two-engine-framework.md) completely before researching or rating a company. For mid-cap discovery or when the user already has an index core, also read [references/mid-cap-opportunity-framework.md](references/mid-cap-opportunity-framework.md) completely. For volatile growth, turnaround, order-book, capacity-expansion, or catalyst-led companies, also read [references/catalyst-watchlist-framework.md](references/catalyst-watchlist-framework.md) completely. 4. Gather current, dated evidence. Prefer exchange filings, regulatory filings, company results and presentations, conference-call transcripts, and official transaction disclosures. Use secondary sources for context, not as substitutes for material primary facts. Cite sources and distinguish facts, estimates, and inferences. 5. Normalize the economic record. Reconcile reported and diluted per-share figures; remove material one-offs; inspect share-count changes, cyclicality, leverage, capitalized costs, acquisitions, and cash conversion. Never annualize one quarter without explaining seasonality. 6. Select the appropriate valuation lens. Use P/E only when positive normalized EPS represents the business. Use the sector-specific alternatives in the reference when P/E is misleading. 7. Assess the EPS engine, valuation engine, promoter signal, dividends and buybacks, balance-sheet risk, and thesis invalidators. Do not count buybacks twice: their share-count effect belongs in EPS; cash dividends belong in total return. 8. Build bear, base, and bull scenarios over the stated horizon. When P/E is valid, run `scripts/project_returns.py` using a temporary JSON input outside this source repository. Report EPS growth, multiple change, dividend contribution, price CAGR, and annualized total shareholder return separately. Compare each case with the return hurdle and calculate the maximum entry price supported by that case. 9. Assign `Discover`, `Validate`, `Starter`, `Scale/Add`, `Hold`, `Trim`, or `Exit/Archive`. Base the state on prospective return, downside, thesis quality, evidence maturity, confidence, concentration, taxes, and opportunity cost—not acquisition price or past return alone. Map `Discover` and `Validate` to `Watch` in compact action summaries. Do not use `Exit` solely because valuation is rich while the operating thesis remains intact; normally use `Hold/Watch` or `Trim`. 10. For a portfolio, aggregate position weights, sector and factor concentration, balance-sheet and governance risks, liquidity, correlated thesis failures, and the effect of proposed trades. Do not impose a fixed HNI or 80/20 allocation. 11. When opportunity discovery is in scope, build a current mid-cap candidate funnel beyond existing holdings. Use the current regulator/AMFI or exchange classification rather than a stale fixed market-cap threshold. Compare candidates with relevant mid-cap peers and retain only ideas that pass the reference's quality, governance, liquidity, valuation, and evidence gates. A watchlist is preferable to a forced buy. ## Portfolio Mandate and Opportunity Discovery - Treat disclosed index funds as the core exposure. Analyze held large caps and flag exceptional opportunities, but do not let familiar index constituents crowd out mid-cap research or duplicate the core without a clear prospective-return advantage. - Separate `core/index exposure`, `active compounders`, `cyclical opportunities`, and `catalyst-growth/emerging compounders`, and `speculative/research positions`. Evaluate concentration across the combined exposure. - For new-money recommendations, prioritize the best risk-adjusted expected returns, with deliberate coverage of mid caps. Do not prefer a mid cap merely because of its size; require a longer runway, credible per-share earnings compounding, acceptable downside, and sufficient liquidity. - Demand a higher margin of safety or return buffer for mid caps because estimates, liquidity, key-person dependence, customer concentration, and governance risk are usually wider than for large caps. - Identify why an opportunity may be mispriced and what evidence can close that gap. Reject ideas whose base return requires speculative P/E expansion, peak margins, an unfinanced capacity plan, or unsupported market-share assumptions. - Prefer a small, ranked shortlist of deeply researched candidates over a long screen. Give explicit buy zones, staged position-size ranges, monitoring events, and thesis invalidators; do not fill an allocation quota when no candidate clears the hurdle. ## Decision Discipline - Use the user's hurdle rate. If absent, disclose an illustrative hurdle and show how the action changes at nearby hurdles. - Treat an attractive base case as insufficient when the bear case risks permanent capital loss or depends on fragile financing or governance. - Treat high-quality growth at a hostile valuation as a valuation problem, not automatically a business-quality problem. - Distinguish volatility from permanent-capital-loss risk. A volatile company with a measurable, funded growth path may merit a capped starter position and catalyst watch even before it qualifies as a proven compounder. - Reserve `Exit` for a broken thesis, unacceptable governance or financing risk, or a clearly superior use of capital after tax and position-size effects. Use `Trim` when price has outrun plausible earnings but the thesis remains valid. - Treat a low multiple with weak normalized earnings as a possible value trap. - Treat promoter buying as corroborating evidence, never as a substitute for operating evidence or valuation. - Do not equate higher return potential with a smaller market cap. Require prospective per-share return to come primarily from earnings/book-value growth and cash returns, with a neutral or conservative valuation engine in the base case. - State confidence as high, medium, or low and list the evidence that would change the action. ## Output Write portfolio analysis beside the input as `.analysis.md`, except when the input is inside this public skill repository. Keep runtime reports, downloads, and intermediate data in the input directory, a private workspace, or a temporary directory—never in this skill repository. Include: 1. As-of date, horizon, assumptions, data gaps, and source policy. 2. Portfolio summary with concentration risks and prioritized actions. Show disclosed index exposure separately and explain where direct holdings overlap. 3. For each company: - position and acquisition details, with cost basis separated from forward value; - normalized operating record and accounting-quality notes; - EPS-engine rating and supporting evidence; - valuation-engine rating and appropriate valuation method; - bear/base/bull return table; - promoter transaction and pledge assessment; - dividends, buybacks, dilution, and capital allocation; - thesis, catalysts, risks, and explicit invalidators; - action, confidence, position-size implication, and concise rationale; - dated source links. 4. When opportunity discovery is requested, a ranked mid-cap shortlist containing: - current market-cap classification and liquidity evidence; - the differentiated thesis and source of potential mispricing; - EPS-engine, valuation-engine, governance, balance-sheet, and cash-conversion gates; - bear/base/bull returns, buy zone, initial and maximum position-size range; - catalyst path, invalidators, confidence, and dated primary sources; - screened candidates rejected and the decisive reason, to expose selection bias. 5. A final action queue: add, hold, trim, exit, and watch/research. Distinguish changes to existing holdings from new mid-cap ideas. 6. For catalyst-growth names, a watchlist dashboard containing the current state, thesis, dated baseline, next catalyst, operating KPIs, valuation/add zone, scale conditions, trim conditions, hard invalidators, position cap, review date, and source links. State which alerts require one quarter and which require confirmation across two quarters. Use ranges rather than false precision. Never present scenario outputs or article case studies as guarantees.