Earnings RecapSAFE

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: earnings-recap
description: >
  Analyze a company's most recent (or a specified past) earnings report from Yahoo
  Finance data (yfinance): actual vs estimated EPS, surprise size, revenue and margin
  trends, and the stock's price reaction. Use this skill whenever the user asks how
  earnings went — beat or miss, earnings surprise, quarterly results, the
  post-earnings move, or an earnings call recap — including casual references to a
  past report such as "AMZN reported last night" or "how did they do". For an
  upcoming report, use earnings-preview.
---

# Earnings Recap Skill

Generates a post-earnings analysis using Yahoo Finance data via [yfinance](https://github.com/ranaroussi/yfinance). Covers the actual vs estimated numbers, surprise magnitude, stock price reaction, and financial context — a complete picture of what happened.

**Important**: Data is for research and educational purposes only. Not financial advice. yfinance is not affiliated with Yahoo, Inc.

---

## Step 1: Ensure yfinance Is Available

**Current environment status:**

```
!`python3 -c "exec('try:\n import yfinance\n print(\'yfinance \' + yfinance.__version__ + \' installed\')\nexcept Exception:\n print(\'YFINANCE_NOT_INSTALLED\')')"`
```

If `YFINANCE_NOT_INSTALLED`, install it:

```python
import subprocess, sys
subprocess.check_call([sys.executable, "-m", "pip", "install", "-q", "yfinance"])
```

If already installed, skip to the next step.

---

## Step 2: Identify the Ticker and Gather Data

Extract the ticker from the user's request. Fetch all relevant post-earnings data in one script.

```python
import yfinance as yf
import pandas as pd

ticker = yf.Ticker("AAPL")  # replace with actual ticker

# --- Earnings results ---
earnings_dates = ticker.get_earnings_dates(limit=12)  # report timestamps, newest first
earnings_hist = ticker.earnings_history               # last 4 quarters, indexed by fiscal quarter-end, oldest first

# --- Financial statements (about five quarters, newest first) ---
quarterly_income = ticker.quarterly_income_stmt
quarterly_cashflow = ticker.quarterly_cashflow
quarterly_balance = ticker.quarterly_balance_sheet

# --- Context ---
info = ticker.info
news = ticker.news
recommendations = ticker.recommendations
```

### What to extract

| Data Source | Key Fields | Purpose |
|---|---|---|
| `get_earnings_dates()` | Earnings Date, EPS Estimate, Reported EPS, Surprise(%) | Which report, when, and the beat/miss |
| `earnings_history` | epsEstimate, epsActual, epsDifference, surprisePercent | Last four quarters' results by fiscal quarter |
| `quarterly_income_stmt` | TotalRevenue, GrossProfit, OperatingIncome, NetIncome, BasicEPS | Actual financials |
| `history()` | Daily closes around each report | Stock price reaction |
| `info` | currentPrice, marketCap, forwardPE | Current context |
| `news` | Recent headlines | Earnings-related news |

---

## Step 3: Find the Report and Measure the Reaction

`earnings_history` is indexed by fiscal quarter-end, not by announcement date, so take report timing from `get_earnings_dates()`: the most recent report is the newest row with a `Reported EPS`. Its timestamp (US Eastern) sets the reaction window: at or after 16:00 means the company reported after the close; anything earlier means before the open or, occasionally, during the session. If the user asked about a specific quarter, use that row instead.

```python
def earnings_reaction(ticker, report_ts):
    """% move from the last close before the report to the first close after it."""
    daily = ticker.history(start=(report_ts - pd.Timedelta(days=10)).date(),
                           end=(report_ts + pd.Timedelta(days=10)).date())
    closes = daily["Close"]
    days = closes.index.date
    d = report_ts.date()
    if report_ts.hour >= 16:  # reported after the close: report-day close -> next close
        pre, post = closes[days <= d], closes[days > d]
    else:                     # before the open or intraday: prior close -> report-day close
        pre, post = closes[days < d], closes[days >= d]
    if pre.empty or post.empty:
        return None           # the reaction session hasn't closed yet
    return (post.iloc[0] / pre.iloc[-1] - 1) * 100

reported = earnings_dates[earnings_dates["Reported EPS"].notna()]
latest_ts = reported.index[0]
reaction_pct = earnings_reaction(ticker, latest_ts)

# Typical earnings-day move over the prior four reports
prior_moves = [earnings_reaction(ticker, ts) for ts in reported.index[1:5]]
avg_abs_move = pd.Series([abs(m) for m in prior_moves if m is not None]).mean()
```

If `reaction_pct` is `None`, the report came after the most recent close; say the regular-session reaction is still pending (an intraday `history(..., prepost=True)` call shows the after-hours move if the user wants it).

---

## Step 4: Build the Earnings Recap

Cover these areas, leading with the result:

1. **Headline result** — EPS actual vs estimate with the surprise %, revenue with year-over-year growth, and the stock's reaction.
2. **Estimates vs actuals** — EPS estimate, actual, and surprise ($ and %) for the quarter in question.
3. **Quarterly trends** — revenue, gross margin, operating margin, and EPS for the recent quarters, with margins computed from the statements (gross profit / revenue, operating income / revenue). yfinance usually returns about five quarters, so year-over-year growth is available for the latest quarter only (column 0 vs column 4); show sequential change for the others rather than inventing a comparison.
4. **Price reaction** — the move in the reaction session, how it compares with the stock's average absolute earnings move over the prior four reports, and whether the stock has since held, given back, or extended the move.
5. **What changed** — margin direction vs the prior quarter, any shift in the revenue growth trajectory, how this surprise compares with the company's usual pattern, and current analyst sentiment if available.

---

## Step 5: Respon

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.