Gemini Deep Research 1.0.0

Perform complex, long-running research tasks using Gemini Deep Research Agent. Use when asked to research topics requiring multi-source synthesis, competitiv...

Security Scan

The skill's code and instructions match its stated purpose (using a Gemini API key to run a long-running research agent); nothing in the files appears to perform unrelated or hidden actions, but there are a few provenance and minor-scope details you should review before installing.

Purpose & Capability

Name/description, SKILL.md, and the included Python script all consistently implement a Gemini Deep Research client that uses a Gemini API key to create/poll interactions and save results. The requested env var (GEMINI_API_KEY) and the declared primary credential align with the described functionality.

Instruction Scope

Runtime instructions direct running scripts/deep_research.py which only use the GEMINI_API_KEY and optional file_search_store. The optional file-search feature will cause the agent to request access to a named file store (potentially exposing local or cloud file contents) — this is within the skill's advertised capabilities but is a privileged operation the user should explicitly permit. The script writes timestamped output files to disk (output-dir), which is expected behavior but will persist results locally.

Install Mechanism

There is no install spec (instruction-only plus a bundled script). No remote downloads or unusual installers are used. Note: the script depends on the Python requests package but no dependency list is declared; this is an operational (not security) mismatch—script will fail if requests is missing.

Credentials

The skill only requires GEMINI_API_KEY (declared as primaryEnv) which is appropriate for calling the Generative Language endpoint with an API key. No other secrets or unrelated env vars are requested. The key will be sent as the x-goog-api-key header to the Google endpoint, which may permit billing/usage on your account—so protect it accordingly.

Persistence & Privilege

always is false and model invocation is allowed (default). The skill does not request persistent system privileges, modify other skills, or require system-wide configuration changes. It writes its own output files only to the configured output directory.

Assessment

This skill is internally consistent with its description, but review a few points before installing: 1) Confirm the GEMINI_API_KEY you provide is appropriate to use here and that you trust the skill/source—API keys enable usage and billing. 2) Be cautious when using the --file-search-store option: that grants the agent access to named file stores and could expose sensitive documents; only specify stores you trust. 3) The package has no install spec and requires Python + the requests library; run it in an environment where you control dependencies (ideally an isolated virtualenv or container). 4) There is a mismatch in provenance metadata: the registry metadata owner ID shown to you differs from ownerId in _meta.json inside the package—this doesn’t prove maliciousness but reduces provenance confidence; if author authenticity matters, ask the publisher for clarification. 5) If you’re unsure, run the script with a limited/test API key and in an isolated environment first.

Gemini Deep Research

Use Gemini's Deep Research Agent to perform complex, long-running context gathering and synthesis tasks.

Prerequisites

• GEMINI_API_KEY environment variable (from Google AI Studio)
• Note: This does NOT work with Antigravity OAuth tokens. Requires a direct Gemini API key.

How It Works

Deep Research is an agent that:

1. Breaks down complex queries into sub-questions
2. Searches the web systematically
3. Synthesizes findings into comprehensive reports
4. Provides streaming progress updates

Usage

Basic Research

scripts/deep_research.py --query "Research the history of Google TPUs"

Custom Output Format

scripts/deep_research.py --query "Research the competitive landscape of EV batteries" \
  --format "1. Executive Summary\n2. Key Players (include data table)\n3. Supply Chain Risks"

With File Search (optional)

scripts/deep_research.py --query "Compare our 2025 fiscal year report against current public web news" \
  --file-search-store "fileSearchStores/my-store-name"

Stream Progress

scripts/deep_research.py --query "Your research topic" --stream

Output

The script saves results to timestamped files:

• deep-research-YYYY-MM-DD-HH-MM-SS.md - Final report in markdown
• deep-research-YYYY-MM-DD-HH-MM-SS.json - Full interaction metadata

API Details

• Endpoint: https://generativelanguage.googleapis.com/v1beta/interactions
• Agent: deep-research-pro-preview-12-2025
• Auth: x-goog-api-key header (NOT OAuth Bearer token)

Limitations

• Requires Gemini API key (get from Google AI Studio)
• Does NOT work with Antigravity OAuth authentication
• Long-running tasks (minutes to hours depending on complexity)
• May incur API costs depending on your quota