context-driven-testing

Context-Driven Testing is a methodology designed for software quality engineers who prioritize actionable results over rigid, one-size-fits-all processes. This skill enables agents to analyze specific project environments—including safety-critical requirements, startup velocity, resource constraints, and team expertise—to formulate testing strategies that find relevant problems effectively. Instead of adhering to static test plans or bureaucratic 'best practices,' the agent performs a context-aware analysis that dictates the depth of automation, the necessity of documentation, and the selection of appropriate testing heuristics.

• Employs RST (Rapid Software Testing) heuristics such as SFDIPOT (Structure, Function, Data, Interfaces, Platform, Operations, Time) to systematically explore software risks.

• Utilizes multiple oracles, including consistency with history, documentation, and user expectations, to validate software behavior.

• Integrates with automated agent fleets for tasks like risk-based test generation, security scanning, and performance validation, prioritizing critical paths based on current project maturity.

• Facilitates adaptive decision-making: choosing light exploratory sessions for fast-paced development or rigorous, documented protocols for regulated industries like medical device software.

• Coordinates with specialized agents within the AQE (Agentic Quality Engineering) framework to scale context-driven thinking without sacrificing human-led critical judgment.

• Use this skill when initiating new project testing strategies or when existing processes feel misaligned with current project risks.

• Inputs typically include project metadata such as stage (e.g., greenfield, maintenance), constraints (timeline, budget), and identified technical risks (e.g., payment security, high volume).

• Expected outputs include an adaptive testing strategy document, a prioritized list of exploratory test tours, and recommendations for automated suite depth.

• The agent functions best when paired with exploratory testing sessions, where discoveries are documented as learning logs rather than rigid script executions.

• Constraints include the necessity for human oversight in defining the initial 'context'—the more accurately you define the project reality, the more precise the agent's strategy adaptation will be.