apify-financial-osint

What this skill does

# Financial OSINT — Social Listening

Discover and quantify what the internet is saying about portfolio companies. Three verified Apify Actors only — Reddit (sentiment + threaded discussion), Twitter/X (real-time mentions, crisis monitoring), Trustpilot (customer satisfaction). All actors verified against real demo data with ≥98% success rate.

## Prerequisites

- Apify access — preferred: `apify` CLI (`npm install -g apify-cli && apify login`); fallback: Apify MCP connector (`call-actor` tool). CLI is faster and preferred when both are available.
- Companies data at `${CLAUDE_PLUGIN_ROOT}/data/companies.json` (read fields: `queries.reddit`, `queries.twitter`, `trustpilot_urls`, `identifiers.ticker`)
- Per-company routing pre-computed at `${CLAUDE_PLUGIN_ROOT}/skills/apify-financial-osint/data/osint-targets.json`

`${CLAUDE_PLUGIN_ROOT}` is the plugin's root directory (where `.claude-plugin/` lives). It is resolved automatically by Claude Code when the plugin is installed, or set to the `--plugin-dir` path during development.

## Workflow checklist

Copy this and tick boxes as you progress:

```
Task Progress:
- [ ] Step 0: Verify Apify access — try `apify --version && apify info`; if unavailable, check for `call-actor` MCP tool; if neither, tell user to install apify CLI or Apify MCP connector
- [ ] Step 1: Pick actor(s) by signal type — see "Choose Actor by Signal" table
- [ ] Step 2: Build input — read data/osint-targets.json or construct from data/companies.json
- [ ] Step 3: Run actor via apify CLI
- [ ] Step 4: Output — present top results with sentiment + engagement signals
```

## Constraints

### Allowed Apify Actors (exhaustive — do NOT use others)

| Actor | Purpose | Cost | Success rate |
|-------|---------|------|--------------|
| `fatihtahta/reddit-scraper-search-fast` | Reddit sentiment, acquisition reactions, brand perception | $1.49 / 1k results | 98.4% (40,787 runs/30d) |
| `kaitoeasyapi/twitter-x-data-tweet-scraper-pay-per-result-cheapest` | Real-time mentions, crisis monitoring, dealflow signals | $0.25 / 1k tweets | 99.7% (4.3/5, 58 reviews) |
| `getwally.net/trustpilot-reviews-scraper` | Service quality, complaint patterns (telcos, e-commerce, banks) | $3.00 / 1k results | verified working |

Do NOT use any other actor. Do NOT use WebSearch, WebFetch, or browser tools.

### Choose Actor by Signal

| If you need | Use Actor | When NOT to use |
|---|---|---|
| Sentiment / discussion threads / reactions to corporate events | `fatihtahta/reddit-scraper-search-fast` | If company has no consumer base (B2B fintech, biotech) — expect <5 posts |
| Real-time mentions / crisis signals / dealflow chatter | `kaitoeasyapi/twitter-x-data-tweet-scraper-pay-per-result-cheapest` | If you need >1 week historical depth — Twitter API limits |
| Customer satisfaction / service quality complaints | `getwally.net/trustpilot-reviews-scraper` | If company has no Trustpilot page (B2B, holding companies) — see verified URL list in `reference/osint-actor-schemas.md` Section 3 |

## Pipeline

### Step 1: Pick actor(s)

For portfolio companies, look up the company in [`data/osint-targets.json`](data/osint-targets.json) — it pre-computes which actors to run with templated inputs. Routing rule (mirrors how the file was built):

- `queries.reddit` non-empty → run Reddit actor
- `queries.twitter` non-empty → run Twitter actor (always set for tracked companies)
- `trustpilot_urls` non-empty → run Trustpilot actor

For ad-hoc / non-portfolio targets, construct input from scratch (see Step 2).

### Step 2: Build input

#### Reddit input (key fields)

| Field | Type | Default | Notes |
|---|---|---|---|
| `queries` | array of string | required (one of queries / urls / subredditName) | Global Reddit-wide search terms. |
| `maxPosts` | integer | **50000** (!) | **ALWAYS set explicitly** — typical 30-50 for scans, 100-200 for deep-dives. |
| `scrapeComments` | boolean | `false` | Set `true` to extract threaded discussion. |
| `maxComments` | integer | 50000 (!) | Only used when `scrapeComments: true`. Typical 5–10. |
| `sort` | enum | `"relevance"` | One of `relevance`, `hot`, `top`, `new`, `comments`. (NOT `rising` / `best`.) |
| `timeframe` | enum | `"all"` | One of `all`, `year`, `month`, `week`, `day`, `hour`. Must be >= dateFrom–dateTo range. |
| `dateFrom` | string | — | `YYYY-MM-DD`. Post-fetch filter: keep posts from this date onward. |
| `dateTo` | string | — | `YYYY-MM-DD`. Post-fetch filter: keep posts up to this date. |

Example:

```bash
apify call fatihtahta/reddit-scraper-search-fast \
  --input '{"queries":["InPost FedEx acquisition"],"maxPosts":50,"scrapeComments":true,"maxComments":10,"sort":"relevance","timeframe":"month"}' \
  --user-agent apify-awesome-skills/apify-financial-osint
```

#### Twitter/X input (key fields)

| Field | Type | Default | Notes |
|---|---|---|---|
| `twitterContent` | string | — | One of `twitterContent` / `tweetIDs` / `searchTerms`. Twitter advanced-search syntax (`OR`, `-`, `from:`, `since:`). |
| `tweetIDs` | array of string | — | **Plural** — not `tweetId`. |
| `searchTerms` | array of string | — | Each term gets `maxItems` results independently. |
| `maxItems` | integer | 200 | **REQUIRED** — actor fails without it. Pay-per-result. |
| `queryType` | enum | `"Latest"` | One of `Latest`, `Top`, `Photos`, `Videos`. |
| `lang` | string | `"en"` | ISO 639-1. Set `cs`/`pl`/`hu`/`bg`/`sk`/`tr` for single-country B2C; omit for multilingual. |
| `since` / `until` | string | — | Format: `YYYY-MM-DD_HH:MM:SS_UTC` (NOT ISO 8601). |
| `filter:news` / `filter:media` / `min_faves` / `min_retweets` | various | — | Engagement / content filters. |

Example:

```bash
apify call kaitoeasyapi/twitter-x-data-tweet-scraper-pay-per-result-cheapest \
  --input '{"twitterContent":"InPost FedEx acquisition OR INPST","maxItems":100,"queryType":"Latest","since":"2026-01-01_00:00:00_UTC","filter:news":true}' \
  --user-agent apify-awesome-skills/apify-financial-osint
```

#### Trustpilot input (only 2 fields exist!)

| Field | Type | Required | Notes |
|---|---|---|---|
| `startUrls` | array of `{"url": "..."}` objects | Yes | **NOT plain strings** — array of objects. |
| `limit` | integer | No (default 1000) | Set lower to control cost ($3/1k). |

Example:

```bash
apify call getwally.net/trustpilot-reviews-scraper \
  --input '{"startUrls":[{"url":"https://www.trustpilot.com/review/inpost.pl"}],"limit":50}' \
  --user-agent apify-awesome-skills/apify-financial-osint
```

Older docs reference fields like `maxItems`, `includeStatistics`, `includeCompanyDetails`, `onlyNewerThan` — these **do NOT exist** on this actor.

### Step 3: Cost-bound the run

Always cap output before running. Defaults are dangerously high.

| Actor | Cap field | Portfolio scan | Deep-dive |
|---|---|---|---|
| Reddit | `maxPosts` | 30–50 | 100–200 |
| Reddit comments | `maxComments` | 5–10 (only if `scrapeComments: true`) | 20–50 |
| Twitter/X | `maxItems` | 50–100 | 200–500 |
| Trustpilot | `limit` | 30–50 | 100–200 |

Twitter and Trustpilot are pay-per-result — every returned item is billed.

### Step 4: Run

Single example pulling Reddit threads + Twitter mentions for InPost (driven by [`data/osint-targets.json`](data/osint-targets.json)):

```bash
apify call fatihtahta/reddit-scraper-search-fast \
  --input "$(jq -c '.targets[] | select(.company_id=="inpost") | .inputs.reddit' \
    ${CLAUDE_PLUGIN_ROOT}/skills/apify-financial-osint/data/osint-targets.json)" \
  --user-agent apify-awesome-skills/apify-financial-osint \
  --output-dataset > reddit_inpost.json
```

Full per-actor input schema (all 51 Twitter properties, every Reddit enum, every Trustpilot edge case) plus 30+ example invocations: [reference/osint-actor-schemas.md](reference/osint-actor-schemas.md).

### Step 4b: Post-filter Reddit results

Reddit search ignores quotes and matches partial words ("InPost" matches "in post game thread"). After fetching, filter results client-side: keep only posts where any o