Recipe Task

Required Skills [LOAD BEFORE EXECUTION]

1. [LOAD IF NOT ACTIVE] task-analyzer — task analysis and skill selection (rule-advisor handles remaining skill selection)
2. [LOAD IF NOT ACTIVE] llm-friendly-context — clear prompts, handoffs, and generated artifacts

Spawn rule: every spawn_agent call uses fork_turns="none" so the subagent receives only the task message and explicitly provided context.

Task Execution with Metacognitive Analysis

Task: $ARGUMENTS

Mandatory Execution Process

Step 1: Rule Selection via rule-advisor (REQUIRED)

Spawn rule-advisor agent: "Analyze the task and select appropriate rules for: $ARGUMENTS. Provide context about current situation and prerequisites."

ENFORCEMENT: Skipping rule-advisor produces unguided execution with high failure risk.

Step 2: Utilize rule-advisor Output

After receiving rule-advisor's response, proceed with:

1. Understand Task Essence (from metaCognitiveGuidance.taskEssence)

• Focus on fundamental purpose, not surface-level work
• Distinguish between "quick fix" vs "proper solution"

1. Follow Selected Rules (from selectedRules)

• Review each selected rule section
• Apply concrete procedures and guidelines

1. Recognize Past Failures (from metaCognitiveGuidance.pastFailures)

• Apply countermeasures for known failure patterns
• Use suggested alternative approaches

1. Execute First Action (from metaCognitiveGuidance.firstStep)

• Start with recommended action and rationale
• Use suggested tools first

Step 3: Create Task List

Register work steps. Always include: first "Confirm skill constraints", final "Verify skill fidelity".

Break down the task based on rule-advisor's guidance:

• Reflect metaCognitiveGuidance.taskEssence in task descriptions
• Apply metaCognitiveGuidance.firstStep to first task
• Restructure tasks considering warningPatterns
• Set appropriate priorities

Step 4: Execute Implementation

Proceed with task execution following:

• Selected rules from rule-advisor
• Task structure
• Quality standards from applicable rules

Important Notes

• Spawn rule-advisor first: MUST complete metacognitive step before implementation
• Update tasks after rule-advisor: MUST reflect insights in task structure
• Follow metaCognitiveGuidance.firstStep: MUST start with the recommended action
• Monitor warningPatterns: MUST watch for failure patterns throughout execution

Completion Criteria

• [ ] rule-advisor spawned and output received
• [ ] Task essence understood from metaCognitiveGuidance.taskEssence
• [ ] Selected rules reviewed and applied
• [ ] Past failure patterns recognized and countermeasures applied
• [ ] Task list created with skill constraint confirmation and fidelity verification
• [ ] Implementation executed following rule-advisor guidance
• [ ] Warning patterns monitored throughout execution

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: shinpr
• Source: shinpr/codex-workflows
• License: MIT

Install and usage instructions live in the source repository linked above.