Skip to content
Back to skills

Corporate Actions Scanner

ASecurity

Scan for material 8-K corporate actions over a lookback window. For a ticker or watchlist, pulls SEC EDGAR 8-K filings, filters to material items (offerings, private placements, splits, spin-offs, buybacks, M&A, restatements), cross-references Massive news for the headline, and computes T+1 and T+5 price reactions. Complements news-scanner (general) and earnings-drilldown (item 2.02 only) by giving material corporate actions a dedicated surface. Use when running a portfolio review, sanity-che...

  • 7 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added September 6, 2026
developmentgoreactapi

Works with

  • api

Security analysis

A100/100

Pro scans all 5 files and shows the line behind each finding

Scanned September 6, 2026

npx -y skills add rgourley/quant-garage --skill corporate-actions-scanner --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Corporate Actions Scanner?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Corporate Actions Scanner
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/rgourley-corporate-actions-scanner/badge)](https://www.skillsdirectory.com/skills/rgourley-corporate-actions-scanner)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: corporate-actions-scanner
description: Scan for material 8-K corporate actions over a lookback window. For a ticker or watchlist, pulls SEC EDGAR 8-K filings, filters to material items (offerings, private placements, splits, spin-offs, buybacks, M&A, restatements), cross-references Massive news for the headline, and computes T+1 and T+5 price reactions. Complements news-scanner (general) and earnings-drilldown (item 2.02 only) by giving material corporate actions a dedicated surface. Use when running a portfolio review, sanity-checking why a name moved months ago, or asking "what happened to this stock?" that news-scanner's 24-hour default missed.
---

# corporate-actions-scanner

You hand over a ticker or watchlist and a lookback window (default 180
days). The skill pulls every 8-K filing from SEC EDGAR in the window,
classifies each into a materiality bucket, cross-references Massive
news for the headline, and computes the T+1 and T+5 price reactions.

Motivated by a live portfolio review that missed an ALLO public
offering (87.5M shares at $2, 34% dilution) because news-scanner
defaulted to a 24-hour window and the offering was 78 days old.
Nothing else in the toolkit surfaced it.

## When to invoke

- The operator is running a portfolio review and wants retrospective
  corporate-action coverage across the book
- "Why did X move" for an unexplained gap that predates the news-
  scanner default window
- Pre-trade hygiene on a name: is there a pending offering, spin-off,
  or M&A announcement in the last quarter that changes the setup
- The user says "material events", "corporate actions", "8-K scan",
  "offerings", "buybacks", "M&A on my book"

## What you need

- `MASSIVE_API_KEY` for news + reactions (Stocks Basic is sufficient)
- SEC EDGAR is public and requires no key

## What you get back

**Layer 1 canonical JSON** matching [`output-schema.json`](./output-schema.json).
Per-event: filing date, 8-K items, materiality buckets, matched news
headline (when found), T+1 and T+5 close-to-close reactions.

**Layer 2 rendered stream** ranked by |T+5 reaction|. Header per event
with ticker + date + items + buckets, headline underneath when matched,
reaction line, optional URL to the source article. See
[`references/rendering.md`](./references/rendering.md).

## How it works

1. **Fetch ticker metadata** for the CIK (SEC EDGAR is CIK-indexed).
2. **Pull 8-K filings** from `data.sec.gov/submissions/CIK{CIK}.json`,
   filter to the lookback window, drop non-material items.
3. **Fetch news window** via `/v2/reference/news` for the same window,
   used to cross-reference the 8-K with a headline.
4. **Pull daily aggs** covering the filing dates + a T+5 buffer.
5. **Match news to filing** within +/- 2 days, preferring same-day and
   the article with the most flavor-keyword hits.
6. **Detect flavor** (public_offering, atm_offering, private_placement,
   share_repurchase, special_dividend, stock_split, spin_off,
   acquisition_announcement, acquisition_target, restatement) via a
   curated keyword list in the news title + description.
7. **Compute T+1 and T+5 reactions** close-to-close from the filing
   date.
8. **Dedupe** filings with the same ticker, date, and item-bucket set
   (SEC sometimes emits duplicate 8-Ks for the same corporate action).

## Item taxonomy

Material items surfaced by default (see `ITEM_TAXONOMY` in the
implementation): 1.01, 1.02, 1.03, 2.01, 2.03, 2.04, 2.05, 2.06, 3.01,
3.02, 3.03, 4.01, 4.02, 5.01, 5.03, 7.01, 8.01.

Excluded from default surface (routine 8-Ks that outnumber material
ones ~10:1): 2.02 (earnings — earnings-drilldown handles), 5.02
(officer/director change), 5.05 (ethics code), 5.07 (vote results),
5.08 (shareholder nomination). Pass `material_only=False` to include
these.

## Endpoints used

- `GET https://data.sec.gov/submissions/CIK{CIK}.json` (SEC EDGAR)
- `GET /v2/reference/news?ticker={T}&published_utc.gte={from}` (Massive)
- `GET /v2/aggs/ticker/{T}/range/1/day/{from}/{to}` (Massive)
- `GET /v3/reference/tickers/{T}` (Massive, for CIK)

## Doesn't handle (yet)

- **Match ambiguity on multi-hit windows.** When multiple news articles
  hit within the +/- 2 day window and all contain flavor keywords, the
  match picks the closest by date. Rare but real.
- **Private-placement lag.** Some material actions surface in 8-K days
  or weeks after they happened; the tool reports the filing date, not
  the effective date.

Files in this skill

  • README.md2.3 KB
  • SKILL.md4.3 KB
  • output-schema.json4 KB
  • references/rendering.md1.8 KB
  • requires.yml1.1 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.