Patterns for wrapping any agent with RAG context from Qdrant. Use to add persistent memory to imported or external agents.
Augments any agent with persistent memory by retrieving relevant context from Qdrant before task execution. Use this pattern when delegating to agents for implementation, debugging, or architecture work where historical decisions and patterns would be valuable.
/plugin marketplace add mindmorass/reflex/plugin install reflex@mindmorass-reflexThis skill inherits all available tools. When active, it can use any tool Claude has access to.
Patterns for augmenting any agent with Qdrant context retrieval.
To wrap an agent with RAG:
Use rag-proxy agent:
Target: {agent-to-wrap}
Task: {the task}
If you need custom control, follow this pattern:
Tool: qdrant-find
Query: {key terms from task}
## Retrieved Context
### Source: {metadata.source}
Harvested: {metadata.harvested_at}
Type: {metadata.type}
{document content}
---
{context blocks}
## Task
{original task}
---
Note: Above context is from stored knowledge. Verify if needed.
Tool: Task
Agent: {target-agent}
Prompt: {enriched prompt}
# Context from Stored Knowledge
The following relevant information was retrieved from project memory:
{{#each contexts}}
## From Qdrant
**Source:** {{metadata.source}}
**Harvested:** {{metadata.harvested_at}}
{{content}}
---
{{/each}}
# Your Task
{{original_task}}
---
**Note:** The context above comes from previously harvested research.
Use it if relevant, but verify currency for time-sensitive information.
The `harvested_at` dates indicate when the content was stored.
Not all tasks need RAG. Skip for:
| Task Type | Wrap? | Reason |
|---|---|---|
| Fresh research | No | Need current, not cached data |
| Simple edits | No | Context not needed |
| RAG-aware agents | No | Already query Qdrant |
| Implementation | Yes | Benefit from patterns, decisions |
| Debugging | Yes | Previous solutions may help |
| Architecture | Yes | Decisions and constraints matter |
Map agent types to useful query topics:
| Agent Type | Query Topics |
|---|---|
| frontend-developer | react, design system, components |
| backend-architect | api, architecture, decisions |
| security-auditor | security, authentication, vulnerabilities |
| devops | infrastructure, terraform, deployment |
| tester | testing, coverage, quality |
After the target agent completes:
Tool: qdrant-store
Information: "<valuable findings>"
Metadata:
source: "agent-output"
type: "generated"
harvested_at: "<ISO date>"
tags: "<relevant,keywords>"
| Scenario | Action |
|---|---|
| Empty query results | Proceed without context |
| Qdrant unavailable | Fall back to unwrapped delegation |
| Target agent fails | Report error, don't retry with less context |
This skill should be used when the user asks to "create a slash command", "add a command", "write a custom command", "define command arguments", "use command frontmatter", "organize commands", "create command with file references", "interactive command", "use AskUserQuestion in command", or needs guidance on slash command structure, YAML frontmatter fields, dynamic arguments, bash execution in commands, user interaction patterns, or command development best practices for Claude Code.
This skill should be used when the user asks to "create an agent", "add an agent", "write a subagent", "agent frontmatter", "when to use description", "agent examples", "agent tools", "agent colors", "autonomous agent", or needs guidance on agent structure, system prompts, triggering conditions, or agent development best practices for Claude Code plugins.
This skill should be used when the user asks to "create a hook", "add a PreToolUse/PostToolUse/Stop hook", "validate tool use", "implement prompt-based hooks", "use ${CLAUDE_PLUGIN_ROOT}", "set up event-driven automation", "block dangerous commands", or mentions hook events (PreToolUse, PostToolUse, Stop, SubagentStop, SessionStart, SessionEnd, UserPromptSubmit, PreCompact, Notification). Provides comprehensive guidance for creating and implementing Claude Code plugin hooks with focus on advanced prompt-based hooks API.