Gemini 4 Argon Prompt Generator

Synthesise the following five research papers on [topic]. For each paper extract: title, year, sample size, key finding, and methodology. Then: (1) identify the three most important points of agreement across studies, (2) identify every material contradiction and hypothesise the most plausible methodological reason for each, (3) rate the overall evidence strength on a five-point scale with justification, (4) write a 300-word evidence-based summary suitable for a specialist audience. Flag every claim that rests on a single study. [Paste paper abstracts or full texts]

Audit this entire codebase for security vulnerabilities. For every issue found, provide: (1) the file path and line number, (2) the vulnerability class (OWASP Top 10, CWE identifier if applicable), (3) the exact attack vector — show the specific input or state that triggers it, (4) the severity rating (Critical/High/Medium/Low) with justification, (5) a minimal, targeted fix with the corrected code. After the issue list, write an executive summary of the three most critical risks and the total attack surface exposed. Do not report style issues or non-security findings. [Paste codebase or repository link]

Review this contract as a general-purpose legal analyst. Identify and explain in plain language: (1) every clause that transfers risk to the party I represent — rank them by severity, (2) any unusual or one-sided provisions not found in standard market-practice agreements, (3) any ambiguities that could be exploited by the counterparty, (4) the three questions I should ask a lawyer before signing, (5) a redline summary of the five changes most worth negotiating. Frame everything as questions and considerations rather than legal advice. [Paste contract text]

Analyse this [report / annual filing / whitepaper] in full. I need: (1) a 200-word executive summary covering the three most important claims, (2) a structured table of every key data point — source page, figure, and your confidence in its accuracy, (3) all forward-looking statements and the assumptions they depend on, (4) any inconsistencies between the narrative and the data tables, (5) a list of 8 questions a rigorous reviewer would ask the authors. Keep section headers. [Paste document — up to 1 million tokens]

Design a complete agentic workflow for the following task: [describe the goal, e.g. 'monitor competitor pricing, detect drops above 10%, generate a Slack alert with a recommended response, and log everything to a Google Sheet']. Specify: (1) every tool call the agent needs (API, web search, database, notification), (2) the decision tree that determines which branch executes, (3) where human checkpoints must be inserted before irreversible actions, (4) how the agent should handle each failure mode, (5) a pseudocode implementation in Python using any standard agent framework. Then list the three biggest reliability risks and how to test for them.

Review this system architecture for production readiness at [expected scale, e.g. 10M requests/day]. Evaluate: (1) single points of failure — list every component whose outage would cause downtime, with an estimated blast radius, (2) performance bottlenecks — predict where the system saturates first under load, (3) data consistency risks across distributed components, (4) security boundaries — where trust is assumed but not enforced, (5) operational complexity — which parts will be hardest to debug at 3 AM. Provide a prioritised remediation list, with the cheapest fix listed first. [Paste architecture diagram description or config files]

Conduct a systematic literature review on [research question] using the studies I have provided. Apply PRISMA criteria: (1) screen each study for inclusion/exclusion against these criteria: [list criteria], (2) extract data into a structured table — author, year, design, sample, outcome measures, key results, risk of bias, (3) perform a narrative synthesis grouped by study design, (4) identify the three largest gaps in current evidence, (5) write an abstract-style summary under 250 words. Mark any studies where you would rate the risk of bias as high and explain why. [Paste study texts or abstracts]

Write a 700-word internal policy memo recommending [specific policy change] to [audience, e.g. 'the CFO of a 500-person SaaS company']. Structure: (1) executive summary in three sentences, (2) problem statement with one quantified metric, (3) recommended action with three supporting arguments ranked by strength, (4) a dedicated section pre-empting the two most likely objections — steelman the objection, then rebut it, (5) implementation steps with owners and 30/60/90-day milestones, (6) risks and mitigation. Tone: confident and direct, no management jargon.

Audit this financial model for a [funding round / board presentation / investor due diligence]. Check: (1) every formula for calculation errors — list each error with the cell reference, what it calculates, and what it should calculate, (2) assumptions that are never stated explicitly but drive the outcome, (3) sensitivity analysis — which three input variables swing the output by more than 20% if moved by 10%, (4) whether the unit economics hold at the projected scale, (5) the three weakest assumptions an investor will challenge first. Provide a short 'clean' version of the key metrics table with errors corrected. [Paste model data or describe assumptions]

Debug this system failure. I will give you: the error output, the relevant code, the last 500 lines of logs before the failure, and the git diff of changes deployed in the last 24 hours. Work through this in order: (1) identify every plausible cause ranked by likelihood, (2) for the top cause, write the diagnostic command I should run and predict exactly what its output will look like if the hypothesis is correct, (3) write the minimal code fix, (4) write a regression test that would have caught this before deployment, (5) identify whether this failure class can recur elsewhere in the codebase. [Paste error, code, logs, diff]

Review the following product feature / data handling process against [GDPR / HIPAA / SOC 2 / specify regulation]. For each applicable requirement: (1) state whether the current implementation is compliant, partially compliant, or non-compliant, (2) cite the specific article or control number, (3) for any gap, describe the minimum change required to achieve compliance and estimate implementation effort (hours), (4) flag any area where you are uncertain due to regulatory ambiguity and recommend specialist counsel. Produce a compliance matrix table as the final output. [Paste implementation description or code]

Draft a research proposal for [funding body / institution] on the topic of [research question]. Include: (1) background and significance — why this question matters now (200 words), (2) research objectives as three specific, measurable aims, (3) methodology — study design, data sources, analysis plan (300 words), (4) expected outcomes and impact, (5) a realistic timeline as a Gantt-style table across 24 months, (6) anticipated challenges and mitigation strategies. Match the tone and structure to a [NSF / NIH / EU Horizon / Wellcome Trust] grant application. Do not overstate significance; a credible proposal is the goal.

Build a competitive intelligence brief on [company or product]. Using only publicly available information, cover: (1) product positioning and target customer — how they describe themselves vs how customers describe them, (2) pricing model and unit economics (estimate if not public), (3) key technical differentiators — what they can do that competitors cannot, (4) known weaknesses from customer reviews, job postings, and product gaps, (5) strategic signals from recent hires, funding, partnerships, and product announcements in the last 6 months, (6) three actionable implications for our product strategy. Flag every inference that is not directly confirmed by a source.

Design a production-grade data pipeline for the following requirements: [describe: data sources, volume, latency SLA, output destination, transformation logic]. Specify: (1) the recommended architecture with component names and the data flow between them, (2) the technology stack with justification for each choice against the alternatives, (3) the schema for the input and output data with field-level types and nullability, (4) error handling strategy — what happens when upstream data is late, malformed, or missing entirely, (5) an observability plan — the three metrics and two alerts that would catch 95% of production failures. Include a Python skeleton for the core transformation step.

Analyse these patent documents for [technology area]. For each patent provide: title, assignee, priority date, claim scope (one sentence), and the key independent claim in plain language. Then: (1) map the IP landscape — which companies control which sub-areas, (2) identify white spaces — technically logical sub-areas with no current protection, (3) flag any of my proposed features that appear to read on existing claims, (4) list the three patents most relevant to a freedom-to-operate analysis I should have a patent attorney review. [Paste patent texts or numbers]

Write a structured 360-degree performance review for [role] based on the following notes from their manager, peers, and direct reports. [Paste notes]. Structure: (1) three concrete strengths with a specific example for each, (2) two development areas — describe the behavior observed, the impact it has on the team, and one actionable way to improve, (3) overall performance rating with justification, (4) three measurable goals for the next review cycle. Tone: direct, fair, and constructive. Do not soften critical feedback to the point of obscuring it — a useful review is honest.

Critique this scientific paper as if you were a peer reviewer for [Nature / Science / NEJM / specify journal]. Evaluate: (1) research question — is it well-defined, novel, and important? (2) methodology — is the study design appropriate? Are controls adequate? Are sample sizes justified? (3) statistical analysis — are the tests appropriate? Are p-values, confidence intervals, and effect sizes reported correctly? (4) conclusions — are claims supported by the data, or overstated? (5) reproducibility — is there enough information to replicate the study? Provide a structured review with Accept / Minor Revision / Major Revision / Reject recommendation and three mandatory revisions if not accepting. [Paste paper]

Read this entire codebase and produce a refactoring plan. I need: (1) a map of every module, its purpose, and its dependencies — flag circular dependencies, (2) the five areas with the highest technical debt, ranked by the risk they introduce and the cost to fix, (3) a migration path for the highest-risk area — ordered steps, breaking change risks at each step, and the test strategy, (4) a list of every function over 100 lines that should be decomposed, with a suggested decomposition for the longest one, (5) the single change that would have the highest risk-reduction-to-effort ratio. [Paste codebase or repository structure]

Run a scenario planning exercise for [organisation or product] over a 3-year horizon. Define and analyse four distinct scenarios — one optimistic, one pessimistic, one technology-disruption, one regulatory-shift. For each scenario: (1) describe the macro conditions that lead to it, (2) identify the three biggest threats and three biggest opportunities for us specifically, (3) propose the one strategic move that would be most valuable if this scenario materialises, (4) assess which capabilities we would need to build now to maintain optionality across all four scenarios. Present the output as a 2×2 scenario matrix plus a 'no-regrets' actions section at the end.

Analyse the attached dataset and produce a complete analytical report. Include: (1) a data quality assessment — null rates, outliers, distribution skew per column, (2) a summary statistics table for every numeric field, (3) the three most statistically significant correlations and a causal hypothesis for each, (4) the single most actionable insight a non-technical executive could act on this week, (5) two charts I should create (describe the chart type, axes, and what pattern it will reveal), (6) the three highest-confidence predictions you can make from this data with a confidence range. Flag any conclusion that rests on fewer than 30 data points. [Attach CSV or paste data]