Overview

Task Direction Approval prevents unauthorized deviations from the original requirements by requiring explicit user consent before changing direction. It helps teams surface the root cause, present 2–3 viable options with trade-offs, and secure explicit approval before proceeding, ensuring decisions are auditable and aligned with goals.

How This Skill Works

When a deviation is detected (switching libraries/tools, replacing automation with a manual workaround, or altering architecture), the skill analyzes the root cause with evidence. It then presents 2–3 options with trade-offs (speed, risk, maintenance, correctness) and prompts the user to choose before any change is implemented. If approval is not given, no changes occur, and investigation-mode can be used for repeated failures or unclear situations.

When to Use It

• Switching original tech or libraries to alternatives
• Replacing automation with a manual workaround
• Changing architecture or design patterns
• Delivering results different from what was requested
• Repeated failures or unclear situations (investigation-mode)

Quick Start

1. Step 1: Explain the failure root cause clearly (what failed, why it failed, evidence).
2. Step 2: Present 2–3 viable options with trade-offs (speed, risk, maintenance, correctness).
3. Step 3: Ask for explicit user choice and proceed only after approval.

Best Practices

• Explain the root cause clearly with evidence (what failed, why it failed).
• Present 2–3 viable options with trade-offs (speed, risk, maintenance, correctness).
• Document pros/cons and alignment with original requirements.
• Ask for explicit user choice and obtain confirmation before proceeding.
• Use investigation-mode when failures are repetitive or the situation is unclear.

Example Use Cases

• Root cause: GraphQL Codegen failed due to authentication error. Options: 1) Add authentication headers, 2) Try a different endpoint, 3) Download schema file directly. Which approach would you prefer?
• Root cause: Migration risk identified. Options: 1) Roll back to the previous schema, 2) Introduce a feature flag for gradual rollout, 3) Run a canary migration with monitoring. Which plan should we adopt?
• Root cause: UI library upgrade may affect visuals. Options: 1) Incremental refactor of affected components, 2) Maintain parallel UIs during migration, 3) Replace only non-critical screens. Which path do you choose?
• Root cause: Automation script produced inconsistent results. Options: 1) Add additional validation checks, 2) Simplify or refactor logic, 3) Use manual fallback process. Which option is acceptable?
• Root cause: Performance regression after architectural change. Options: 1) Revert to original architecture, 2) Optimize bottlenecks, 3) Introduce caching or batching. Which approach should proceed?

Frequently Asked Questions