Keyword Research

Multi-source keyword intelligence and autocomplete research. Fetches real-time suggestions from Google, YouTube, Amazon, and DuckDuckGo — no API key required...

Security Scan

The skill’s code, instructions, and requirements are consistent with a local keyword-autocomplete fetcher that queries public suggestion endpoints; it requests no credentials or installs and appears to do what it says.

Purpose & Capability

The name/description promise (multi-source autocomplete suggestions) matches the provided SKILL.md and the script. The script queries Google, YouTube, Amazon, DuckDuckGo and Bing suggestion endpoints — exactly the advertised sources. No unrelated credentials, binaries, or installs are requested.

Instruction Scope

SKILL.md instructs running the included Python script with flags. It does not instruct the agent to read arbitrary local files or hidden credentials. The runtime behavior (HTTP requests to public suggestion APIs, optional expand/dedup/output modes) is within the stated scope. Note: running the script transmits whatever seed keyword you provide to third-party suggestion endpoints (privacy consideration).

Install Mechanism

No install spec; this is instruction + included Python script that uses only the standard library. No downloads from external URLs or package registries occur during install. Risk from installation is low.

Credentials

The skill requests no environment variables or credentials (proportionate). However, it sends user-supplied keywords over the network to multiple third-party endpoints — avoid submitting sensitive data as input. There is no hidden credential access.

Persistence & Privilege

always:false and default agent invocation are used. The skill does not request persistent installation, modify other skills, or write agent/system-wide config. It will only run when invoked.

Assessment

This skill appears to be what it claims: a small Python script that queries public autocomplete endpoints. Before using it: (1) review the included scripts/fetch_suggestions.py yourself (it is short and uses only the stdlib), (2) do not pass any sensitive or private strings as the seed keyword — those values are sent to third-party services, (3) run it locally in an environment you control (or sandbox) if you have concerns, (4) be aware of provider rate limits, possible captchas or blocking if you run large batches, and potential Terms-of-Service issues for scraping autocomplete endpoints, and (5) if you plan to automate via an agent, remember that the agent will send whatever keywords it queries to those external endpoints; restrict autonomous runs accordingly.

Keyword Intelligence

Multi-source autocomplete fetcher — no API keys needed. Pulls real-time suggestions from Google, YouTube, Amazon, and DuckDuckGo.

Quick Usage

# All sources, default language (de)
python3 scripts/fetch_suggestions.py "keyword"

# Specific sources
python3 scripts/fetch_suggestions.py "keyword" --sources google,youtube

# English / US region
python3 scripts/fetch_suggestions.py "keyword" --lang en --region us

# Expand mode: fetch suggestions of suggestions (10x more keywords)
python3 scripts/fetch_suggestions.py "keyword" --sources google --expand

# JSON output (for piping or further processing)
python3 scripts/fetch_suggestions.py "keyword" --json

Options

Flag	Default	Description
--sources	all	all or comma-separated: google, youtube, amazon, ddg
--lang	de	Language code: de, en, tr, fr, es, ...
--region	de	Region: de, us, tr, gb, ...
--expand	off	Fetches 2nd-level suggestions from Google (base keywords → ~10x results)
--json	off	Outputs JSON instead of formatted text

Sources

• Google — Broadest coverage, best for general web search intent
• YouTube — Video content ideas, tutorials, how-to queries
• Amazon — Product/buying intent keywords (works best for product niches)
• DuckDuckGo — Privacy-focused users, tech/dev audience

Workflow

1. Start with a seed keyword and --sources all
2. Identify which source is most relevant for the use case
3. Use --expand on the most promising source for deeper research
4. Export with --json to process or display the results

Notes

• No rate limits enforced, but add delays for large batch jobs (the script adds 0.2s between expand calls)
• Amazon suggestions may be empty for non-product keywords — expected behavior
• YouTube returns fewer results for niche/regional keywords
• --expand only works on Google (most reliable for 2nd-level fetching)