recipe-task
GitHub基于元认知分析的通用任务执行流程，通过调用规则顾问选择策略、分析任务本质并规避历史失败模式，按依赖顺序有序执行。
Trigger Scenarios
Install
npx skills add shinpr/claude-code-workflows --skill recipe-task -g -y
SKILL.md
Frontmatter
{
"name": "recipe-task",
"description": "Execute tasks following appropriate rules with rule-advisor metacognition",
"disable-model-invocation": true
}
Explicit User Instruction: The user explicitly instructs and authorizes every subagent call named in this recipe. Execute each applicable call when its prerequisites are met.
Execute Skill: llm-friendly-context before writing Agent prompts, handoffs, or generated artifacts. Execute Skill: subagents-orchestration-guide before making workflow decisions, invoking agents, or resolving findings.
Task Execution with Metacognitive Analysis
Task: $ARGUMENTS
Mandatory Execution Process
Step 1: Rule Selection via rule-advisor (REQUIRED)
Invoke rule-advisor using Agent tool:
subagent_type: "dev-workflows:rule-advisor"description: "Rule selection"prompt: "Task: $ARGUMENTS. Select appropriate rules and perform metacognitive analysis."
Step 2: Utilize rule-advisor Output
After receiving rule-advisor's JSON response, proceed with:
-
Understand Task Essence (from
taskAnalysis.essence)- Focus on fundamental purpose, not surface-level work
- Distinguish between "quick fix" vs "proper solution"
-
Follow Selected Rules (from
selectedRules)- Execute each selected skill by its
skillname and read it completely - Apply the named sections in the context of the complete skill
- Execute each selected skill by its
-
Recognize Past Failures (from
metaCognitiveGuidance.pastFailures)- Apply countermeasures for known failure patterns
- Use suggested alternative approaches
-
Execute First Action (from
metaCognitiveGuidance.firstStep)- Start with recommended action
- Use suggested tools first
Step 3: Bind the Execution Sequence
Before implementation, derive the smallest dependency-ordered sequence required by the rule-advisor result. Its first gate applies and maps the selected rules; its final gate verifies those rules and the requested outcome. Execute one gate at a time, advancing only when its required evidence exists. Add or reorder a gate only when new evidence changes a dependency or completion condition.
Step 4: Execute Implementation
Proceed with task execution following:
- Start with
metaCognitiveGuidance.firstStepaction from rule-advisor - Selected rules from rule-advisor
- Dependency-ordered execution sequence from Step 3
- Quality standards from the selected skills and sections named by rule-advisor
- Monitor warningPatterns flags throughout execution and adjust approach when triggered
Version History
-
e245981
Current 2026-08-19 14:38
移除了对TaskCreate工具工作流的依赖，简化了步骤3的执行逻辑。
-
0d96a63
2026-08-05 22:06
优化了步骤2中规则应用的描述，将'Review each selected rule section'改为'Execute each selected skill by its skill name and read it completely'，强调完整阅读和执行技能而非仅审查。
-
51b7dbc
2026-08-05 01:47
修复：强制执行以结果为导向的编排逻辑
-
d439b50
2026-07-31 02:54
重构了编排任务的跟踪逻辑，将初始任务描述从“映射预加载技能”调整为“选择和映射适用规则”，并将最终验证任务改为“验证选定的规则并报告完成”。
- 66e3b29 2026-07-05 12:01
