leadmagic-skillsSAFE

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: ads-intelligence
description: "LeadMagic competitor ads research — Google Ads, Meta (Facebook/Instagram) Ads, and B2B ads library search plus single-ad details. Use when researching a competitor's ad creatives, messaging, or offers, monitoring ad changes over time, or mining ad copy for positioning and cold-email angles."
license: MIT
compatibility: "Requires network access to api.leadmagic.io or mcp.leadmagic.io."
metadata:
  author: LeadMagic Corporation
  version: "3.0.3"
  homepage: https://leadmagic.io?utm_source=github&utm_medium=skill&utm_campaign=leadmagic-skills
  docs: https://leadmagic.io/docs?utm_source=github&utm_medium=skill&utm_campaign=leadmagic-skills
  github: https://github.com/LeadMagic/leadmagic-skills
  publisher: LeadMagic Corporation
  tags: [leadmagic, ads, google-ads, meta-ads, b2b-ads, competitive-intel]
---

# LeadMagic — Ads intelligence

## Endpoints

| Goal | Endpoint | Credits | Rate/min |
|---|---|---|---|
| Google ads by advertiser | `POST /v1/ads/google-ads-search` | 1 per returned ad | 1,500 |
| Meta (FB/IG) ads | `POST /v1/ads/meta-ads-search` | 1 per returned ad | 1,500 |
| B2B ads library | `POST /v1/ads/b2b-ads-search` | 1 base + 1 per ad | 1,500 |
| One B2B ad, full detail | `POST /v1/ads/b2b-ads-details` | 2 | 1,500 |

Search by `company_domain` (preferred) or company name. **Pass `limit`** to cap creatives returned — it caps spend too. Legacy aliases (`/google/searchads`, `/meta/searchads`, `/b2b/searchads`, `/b2b/ad-details`) still resolve.

```bash
curl -sS -X POST "https://api.leadmagic.io/v1/ads/meta-ads-search" \
  -H "X-API-Key: $LEADMAGIC_API_KEY" -H "Content-Type: application/json" \
  -d '{"company_domain":"competitor.com","limit":10}'
```

## Rules

- Never invent or paraphrase-as-quote ad copy — report only creatives the API returned.
- Details (2) only for creatives worth dissecting; diff against last run first.
- An empty result usually means the advertiser isn't running ads on that network — that's a finding, not an error. Don't retry.

## Outbound patterns

- **Weekly competitor monitor** (recipe 9): 3 searches/competitor (cost depends on returned ads) → diff vs last week → details on new creatives (2) → messaging brief.
- **Angle mining**: competitor's ad promises become your cold-email counter-positioning ("they promise X; here's where X breaks").
- **Warm-account signal**: a target account running B2B ads = active budget + market pressure; stack with hiring signals (`jobs-hiring-intent`) for timing.
- Ads tools are currently paused on the hosted MCP. Use the REST endpoints above when available; do not invent an MCP tool.

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 (3)

If a real credential was exposed, revoke or rotate it with the issuing service. Current-file cleanup does not erase Git history, forks, downloaded releases, or third-party caches. Report suspected exp

Report suspected vulnerabilities privately to [[email protected]](mailto:[email protected]). Share a minimal, sanitized reproduction and affected version. Do not post credentials, access token

An integration with a third-party service does not imply ownership of that service or endorsement by its provider. Repository licensing does not grant access to hosted services, API credentials, or cu

Gates applied: no_behavioural_pass.