Earnings Trade AnalyzerSAFE

Claude Code skills for equity investors and traders — market analysis, technical charting, economic calendars, screeners, and trading strategy development.

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-trade-analyzer
description: Analyze recent post-earnings stocks using a 5-factor scoring system (Gap Size, Pre-Earnings Trend, Volume Trend, MA200 Position, MA50 Position). Scores each stock 0-100 and assigns A/B/C/D grades. Use when user asks about earnings trade analysis, post-earnings momentum screening, earnings gap scoring, or finding best recent earnings reactions.
---

# Earnings Trade Analyzer - Post-Earnings 5-Factor Scoring

Analyze recent post-earnings stocks using a 5-factor weighted scoring system to identify the strongest earnings reactions for potential momentum trades.

## When to Use

- User asks for post-earnings trade analysis or earnings gap screening
- User wants to find the best recent earnings reactions
- User requests earnings momentum scoring or grading
- User asks about post-earnings accumulation day (PEAD) candidates

## Prerequisites

- FMP API key (set `FMP_API_KEY` environment variable or pass `--api-key`)
- Free tier (250 calls/day) is sufficient for default screening (lookback 2 days, top 20)
- Paid tier recommended for larger lookback windows or full screening

## Workflow

### Step 1: Run the Earnings Trade Analyzer

Execute the analyzer script:

```bash
# Default: last 2 days of earnings, top 20 results
python3 skills/earnings-trade-analyzer/scripts/analyze_earnings_trades.py --output-dir reports/

# Custom lookback and market cap filter
python3 skills/earnings-trade-analyzer/scripts/analyze_earnings_trades.py \
  --lookback-days 5 \
  --min-market-cap 1000000000 \
  --top 30 \
  --output-dir reports/

# Deterministic anchor date (America/New_York); the window is anchored on the
# ET calendar date, not the runner's local clock. For reproducible runs/tests.
python3 skills/earnings-trade-analyzer/scripts/analyze_earnings_trades.py \
  --as-of 2026-09-15 \
  --output-dir reports/

# With entry quality filter
python3 skills/earnings-trade-analyzer/scripts/analyze_earnings_trades.py \
  --apply-entry-filter \
  --output-dir reports/
```

#### Degraded endpoint / budget fallback for scheduled reviews

If the analyzer reports a 404, an implausible empty earnings calendar, or exhausts its API-call budget before producing scored candidates during a scheduled after-close/pre-market run, do not report "no earnings reactions" immediately. A clean empty response over a date window containing at least one XNYS session exits 1 with `ZERO_RESULT_REASON=earnings_calendar_empty_with_market_sessions`. If the shared XNYS calendar cannot classify the window, it exits 1 with `ZERO_RESULT_REASON=market_calendar_unavailable`. Budget or daily rate-limit exhaustion during profile fetching exits 1 with `ZERO_RESULT_REASON=profiles_budget_exhausted`. Treat each as a failed run to retry or fall back on, not a quiet day. Only a clean empty response over a zero-session window exits 0 as `ZERO_RESULT_REASON=no_earnings_rows`.

1. First retry once with a narrower liquid-universe configuration so the full 5-factor scorer has a chance to complete, for example:

```bash
python3 skills/earnings-trade-analyzer/scripts/analyze_earnings_trades.py \
  --lookback-days 2 \
  --min-market-cap 5000000000 \
  --top 20 \
  --max-api-calls 600 \
  --output-dir reports/<routine-date>
```

2. If the scored run still returns no candidates or cannot complete, verify the same range through the stable endpoint used by the compatibility shim and clearly label the result as an ungraded fallback:

```bash
curl "https://financialmodelingprep.com/stable/earnings-calendar?from=YYYY-MM-DD&to=YYYY-MM-DD&apikey=$FMP_API_KEY"
```

Then optionally enrich returned US tickers through the analyzer's stable-first FMP client or per-symbol `/stable/quote?symbol=<ticker>` calls to rank by same-day `changesPercentage`, market cap, and liquidity. Use legacy `/api/v3` quote calls only as a legacy-key fallback after stable has failed. Present these as **preliminary / ungraded reactions** because the 5-factor scorer did not run; do not assign A/B/C/D grades from the fallback alone.

**No-candidate output pitfall:** The analyzer may print `Candidates after filtering: 0` / `No candidates found matching criteria.` and exit successfully without writing an `earnings_trade_analyzer_*.json` file. In that case, do not try to run PEAD Mode B from a nonexistent candidate file. Say explicitly that no scored analyzer JSON was produced, run the endpoint/quote enrichment fallback above if the routine needs an earnings section, and label any names as manual-review only. This success-exit path does not cover budget exhaustion during profile fetching: that case exits 1 (`ZERO_RESULT_REASON=profiles_budget_exhausted`) instead.

#### Empty windows and today-only runs

The earnings calendar window is inclusive and uses the `America/New_York`
calendar date from `--as-of` (or the current ET date). A clean provider `[]`
is a benign quiet-window result only when the shared XNYS calendar successfully
counts zero exchange sessions in that exact window, such as a weekend or
holiday. If the window contains an XNYS session, the same clean `[]` exits 1
with `ZERO_RESULT_REASON=earnings_calendar_empty_with_market_sessions` so a
provider drop is not reported as a quiet day. If the XNYS calendar cannot be
queried, the run also exits 1 with
`ZERO_RESULT_REASON=market_calendar_unavailable`.

`--lookback-days 0` is valid and queries exactly the single ET as-of date. Use
it after the relevant announcements have been published (normally after the
session close); an empty response on an XNYS session remains intentionally
fail-closed. A non-empty response whose rows do not carry a `symbol` retains
the separate `ZERO_RESULT_REASON=no_earnings_rows` behavior; that case is not
the literal-empty-list session check above.

### Step 2: Review Results

1. Read the generated JSON and Markdown reports
2. Load `references/scoring_methodology.md` for scoring interpretation context
3. Focus on Grade A and B stocks for actionable setups

### Step 3: Present Anal

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

not all pinned

Secrets in source

none-found

Findings (2)

examples/daily-market-dashboard/.claude/skills

requests

Gates applied: no_behavioural_pass.

[![ModelsAtlas audit: Earnings Trade Analyzer](https://modelsatlas.com/skills/tradermonty/earnings-trade-analyzer/badge.svg)](https://modelsatlas.com/skills/tradermonty/earnings-trade-analyzer/)

Not your listing, or something out of date? Claim or update your listing.