patterns
Common agent patterns and templates for Claude Code. Use when implementing agents to follow proven patterns for proxy mode, TodoWrite integration, and quality checks.
What this skill does
# Agent Patterns
## Proxy Mode Pattern
Enable agents to delegate to external AI models via Claudish.
```xml
<critical_constraints>
<proxy_mode_support>
**FIRST STEP: Check for Proxy Mode Directive**
Before executing, check if the incoming prompt starts with:
```
PROXY_MODE: {model_name}
```
If you see this directive:
1. **Extract model name** (e.g., "x-ai/grok-code-fast-1")
2. **Extract actual task** (everything after PROXY_MODE line)
3. **Construct agent invocation**:
```bash
AGENT_PROMPT="Use the Task tool to launch the '{agent-name}' agent:
{actual_task}"
```
4. **Delegate via Claudish**:
```bash
printf '%s' "$AGENT_PROMPT" | npx claudish --stdin --model {model_name} --quiet --auto-approve
```
5. **Return attributed response**:
```markdown
## {Task Type} via External AI: {model_name}
{EXTERNAL_AI_RESPONSE}
---
*Generated by: {model_name} via Claudish*
```
6. **STOP** - Do not execute locally
**If NO PROXY_MODE directive**: Proceed with normal workflow
</proxy_mode_support>
</critical_constraints>
```
**Key Elements:**
- Check for directive first
- Use `--auto-approve` flag
- Clear attribution in response
- STOP after proxy (don't continue locally)
---
## TodoWrite Integration Pattern
Every agent must track workflow progress.
```xml
<critical_constraints>
<todowrite_requirement>
You MUST use TodoWrite to track your workflow.
**Before starting**, create todo list:
1. Phase 1 description
2. Phase 2 description
3. Phase 3 description
**Update continuously**:
- Mark "in_progress" when starting
- Mark "completed" immediately after finishing
- Keep only ONE task "in_progress" at a time
</todowrite_requirement>
</critical_constraints>
<workflow>
<phase number="1" name="Phase Name">
<step>Initialize TodoWrite with all phases</step>
<step>Mark PHASE 1 as in_progress</step>
<step>... perform work ...</step>
<step>Mark PHASE 1 as completed</step>
<step>Mark PHASE 2 as in_progress</step>
</phase>
</workflow>
```
---
## Quality Checks Pattern (Implementers)
```xml
<implementation_standards>
<quality_checks mandatory="true">
Before presenting code, perform these checks in order:
<check name="formatting" order="1">
<tool>Biome.js</tool>
<command>bun run format</command>
<requirement>Must pass</requirement>
<on_failure>Fix and retry</on_failure>
</check>
<check name="linting" order="2">
<tool>Biome.js</tool>
<command>bun run lint</command>
<requirement>All errors resolved</requirement>
<on_failure>Fix errors, retry</on_failure>
</check>
<check name="type_checking" order="3">
<tool>TypeScript</tool>
<command>bun run typecheck</command>
<requirement>Zero type errors</requirement>
<on_failure>Resolve errors, retry</on_failure>
</check>
<check name="testing" order="4">
<tool>Vitest</tool>
<command>bun test</command>
<requirement>All tests pass</requirement>
<on_failure>Fix failing tests</on_failure>
</check>
</quality_checks>
</implementation_standards>
```
---
## Review Feedback Pattern (Reviewers)
```xml
<review_criteria>
<feedback_format>
## Review: {name}
**Status**: PASS | CONDITIONAL | FAIL
**Reviewer**: {model}
**Issue Summary**:
- CRITICAL: {count}
- HIGH: {count}
- MEDIUM: {count}
- LOW: {count}
### CRITICAL Issues
#### Issue 1: {Title}
- **Category**: YAML | XML | Security | Completeness
- **Description**: What's wrong
- **Impact**: Why it matters
- **Fix**: How to fix it
- **Location**: Section/line reference
### HIGH Priority Issues
[Same format]
### Approval Decision
**Status**: PASS | CONDITIONAL | FAIL
**Rationale**: Why this status
</feedback_format>
</review_criteria>
<approval_criteria>
<status name="PASS">
- 0 CRITICAL issues
- 0-2 HIGH issues
- All core sections present
</status>
<status name="CONDITIONAL">
- 0 CRITICAL issues
- 3-5 HIGH issues
- Core functionality works
</status>
<status name="FAIL">
- 1+ CRITICAL issues
- OR 6+ HIGH issues
- Blocks functionality
</status>
</approval_criteria>
```
---
## Orchestrator Phase Pattern (Commands)
```xml
<phases>
<phase number="1" name="Descriptive Name">
<objective>Clear statement of what this phase achieves</objective>
<steps>
<step>Mark PHASE 1 as in_progress in TodoWrite</step>
<step>Detailed action step</step>
<step>Detailed action step</step>
<step>Mark PHASE 1 as completed</step>
</steps>
<quality_gate>
Exit criteria - what must be true to proceed
</quality_gate>
</phase>
</phases>
<delegation_rules>
<rule scope="design">ALL design → architect agent</rule>
<rule scope="implementation">ALL implementation → developer agent</rule>
<rule scope="review">ALL reviews → reviewer agent</rule>
</delegation_rules>
```
---
## Agent Templates
### Planner Template
```yaml
---
name: {domain}-architect
description: |
Plans {domain} features with comprehensive design.
Examples: (1) "Design X" (2) "Plan Y" (3) "Architect Z"
model: sonnet
color: purple
tools: TodoWrite, Read, Write, Glob, Grep, Bash
---
```
### Implementer Template
```yaml
---
name: {domain}-developer
description: |
Implements {domain} features with quality checks.
Examples: (1) "Create X" (2) "Build Y" (3) "Implement Z"
model: sonnet
color: green
tools: TodoWrite, Read, Write, Edit, Bash, Glob, Grep
---
```
### Reviewer Template
```yaml
---
name: {domain}-reviewer
description: |
Reviews {domain} code for quality and standards.
Examples: (1) "Review X" (2) "Validate Y" (3) "Check Z"
model: sonnet
color: cyan
tools: TodoWrite, Read, Glob, Grep, Bash
---
```
### Orchestrator Template
```yaml
---
description: |
Orchestrates {workflow} with multi-agent coordination.
Workflow: PHASE 1 → PHASE 2 → PHASE 3
allowed-tools: Task, AskUserQuestion, Bash, Read, TodoWrite, Glob, Grep
---
```
Related in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.