Optimize the agent's working environment (project structure, config files, tool availability) to make productive actions obvious and error-prone patterns invisible. Use when setting up a project, auditing a codebase, or reducing friction in development workflows.
Calibrate task complexity so the agent operates at the edge of its abilities — not so easy it produces sloppy output, not so hard it stalls. Use when an agent is generating boilerplate without thinking, overwhelmed by scope, or needs to find the optimal challenge level for a task.
Chain agent operations by linking a new behavior to an existing workflow step. Use when building reliable multi-step pipelines, enforcing quality gates after code generation, or ensuring consistent post-action behaviors like linting, testing, or documentation.
Implement a tracking and feedback loop for agent operations to measure consistency and catch regressions. Use when building CI/CD pipelines, monitoring code quality metrics, tracking task completion rates, or establishing any system where progress must be visible.
Reframe the agent's approach from outcome-chasing to identity-aligned execution. Use when an agent needs to adopt a consistent persona (security engineer, performance hawk, documentation expert) rather than treating each task as an isolated goal.
Create a concrete execution plan specifying exactly what action to take, when to take it, and in what context. Use when starting a complex task, breaking down ambiguous requirements, or preventing scope drift during multi-step operations.
Conduct a periodic audit of the agent's workflows, outputs, and patterns to course-correct and compound improvements. Use after completing a project phase, before starting a new sprint, or when performance has plateaued.
Pair a tedious but necessary agent operation with a high-value operation the agent or user wants. Use when the workflow includes boring but critical steps like testing, documentation, linting, or logging that tend to be skipped.