--- name: Workflow Patterns description: Patterns for output formats, reflection checkpoints, agent references, and self-evaluation shared across agents and commands. version: 1.0.0 --- Provide standardized patterns for output formatting, workflow checkpoints, agent references, and self-evaluation shared across agents and commands. <\/purpose> Standard agent output format with status criteria<\/description> Include in agents for consistent output structure<\/use_case> <\/tool> Standard analysis quality checkpoint for workflow phases<\/description> Include in workflows for consistent quality gates<\/use_case> <\/tool> Standard self-evaluation phase for agents producing reports<\/description> Include in agents for consistent quality assessment<\/use_case> <\/tool> Standard Serena initialization phase for commands<\/description> Include at start of command workflows for Serena initialization<\/use_case> <\/tool> Serena validation elements for reflection checkpoints<\/description> Include in checkpoints before code modifications or after investigations<\/use_case> <\/tool> <\/tools> Standard agent output format with status criteria<\/description> { "status": "success|warning|error", "status_criteria": { "success": "All checks passed, confidence >= 80", "warning": "Minor issues OR confidence 60-79", "error": "Critical issues OR confidence less than 60" }, "confidence": 0, "summary": "Brief summary of results", "metrics": {}, "findings": [], "next_actions": [] } <\/format> <\/output> <\/example> <\/pattern> Standard status criteria for agent output format<\/description> "status_criteria": { "success": "All checks passed, confidence >= 80", "warning": "Minor issues OR confidence 60-79", "error": "Critical issues OR confidence less than 60" } <\/example> <\/pattern> Standard analysis quality checkpoint for workflow phases<\/description> Have I gathered sufficient evidence to proceed?<\/question> Are there gaps in my understanding?<\/question> If confidence less than 70, seek more evidence or ask user<\/threshold> <\/reflection_checkpoint> <\/example> <\/pattern> Standard Serena initialization phase for workflows<\/description> Initialize Serena and check existing patterns<\/objective> 1. Activate Serena project with activate_project<\/step> 2. Check list_memories for relevant patterns<\/step> 3. Load applicable memories with read_memory<\/step> <\/phase> <\/example> <\/pattern> Serena validation for investigation\/analysis checkpoints<\/description> think_about_collected_information<\/tool> After investigation phase completes<\/trigger> <\/serena_validation> <\/example> <\/pattern> Serena validation before code modification operations<\/description> think_about_task_adherence<\/tool> Before code modification<\/trigger> <\/serena_validation> <\/example> <\/pattern> Serena validation before reporting task completion<\/description> think_about_whether_you_are_done<\/tool> Before reporting completion to user<\/trigger> <\/serena_validation> <\/example> <\/pattern> Standard failure handling phase for workflows<\/description> Handle errors and edge cases gracefully<\/objective> If tool call fails: Log error, attempt alternative approach<\/step> If data unavailable: Document gap, proceed with partial analysis<\/step> If contradictory evidence: Flag uncertainty, request user clarification<\/step> <\/phase> <\/example> <\/pattern> Standard agent reference syntax for commands<\/description> <\/agents> Use ref attribute to reference agent defined in agents\/ directory. readonly attribute indicates whether agent can modify files. <\/example> <\/pattern> Standard self-evaluation phase for commands and agents that produce reports<\/description> Brief quality assessment of output<\/objective> Calculate confidence using decision_criteria factors<\/action> Decision criteria evaluation<\/tool> Confidence score<\/output> <\/step> Identify top 1-2 critical issues if confidence below 80<\/action> Gap analysis<\/tool> Issue list<\/output> <\/step> Append self_feedback section to output<\/action> Output formatting<\/tool> Self-feedback section<\/output> <\/step> <\/phase> <\/example> <\/pattern> Standard self-feedback output section for commands that include self-evaluation<\/description> XX\/100 (based on decision_criteria calculation)<\/confidence> - [Critical] Issue description (if any, max 2 total) - [Warning] Issue description (if any) <\/issues> <\/self_feedback> <\/example> <\/pattern> <\/patterns> Use output_format for all agents that return structured results<\/practice> Include reflection_checkpoint at key workflow decision points<\/practice> Include prepare_phase for Serena initialization in commands<\/practice> Add serena_validation to checkpoints before code modifications<\/practice> Add self_evaluate_phase for agents producing reports or recommendations<\/practice> Use failure_handling phase in all workflows<\/practice> Use agent_ref syntax for consistent agent references in commands<\/practice> <\/best_practices> Output status must use standard criteria (success >= 80, warning 60-79, error less than 60)<\/rule> Reflection checkpoints must include confidence threshold<\/rule> Commands must include prepare_phase for Serena initialization<\/rule> Use serena_validation_pre_edit before code modification operations<\/rule> <\/rules> Include failure_handling phase in complex workflows<\/rule> Use self_feedback_output format for self-evaluation results<\/rule> Use agent_ref with readonly attribute for clarity<\/rule> Add serena_validation_investigation after investigation phases<\/rule> <\/rules> Base templates for error escalation, decision criteria, enforcement<\/skill> Parallel execution and timeout configuration<\/skill> Serena MCP operations (memory, symbol search, code navigation)<\/skill> <\/related_skills> Use standard output_status_criteria thresholds<\/must> Include confidence score in all structured outputs<\/must> Define threshold in reflection_checkpoints<\/must> Include prepare_phase in command workflows<\/must> Use serena_validation before code modifications<\/must> Custom status thresholds that differ from standard<\/avoid> Omitting failure_handling in complex workflows<\/avoid> Omitting Serena initialization in commands<\/avoid> <\/constraints>