daa-agent — Claude Code skill: agents that learn and adapt
Builds Dynamic Agentic Architecture agents that shift behavior from performance metrics, pick up experience as they run and pass it to other agents.
npx skills add ruvnet/ruflo --skill daa-agent -a claude-codeWhat daa-agent does
The skill drives five tools on the ruflo MCP server: daa_agent_create, daa_learning_status, daa_performance_metrics, daa_agent_adapt and daa_knowledge_share. The first one takes the starting configuration plus learning parameters. The rest report how adaptation is going, return accuracy and efficiency numbers, fire a manual adaptation from feedback, and push what one agent learned out to the others.
A static agent keeps the configuration you gave it and stays on its own. Here behavior moves over time, learning never really stops, and knowledge is not locked inside a single agent. Calls follow the shape <create|adapt|status> [options].
Nothing runs without the ruflo MCP server connected, because every step lands on one of its tools; Bash is allowed alongside them as a helper. Worth setting up once you run more than one agent and want a good result from one to reach the rest.
What it helps with
- spin up an agent that tunes itself from metrics
- watch adaptation progress across runs
- pull accuracy and efficiency numbers for an agent
- trigger a manual adaptation from feedback
- move what one agent learned to the others
Example prompts for the agent
- Create a DAA agent with an initial configuration and learning parameters
- Show the learning status and performance metrics for this agent
- Adapt the agent from the latest feedback and share the knowledge with the other agents
How to install daa-agent
1Skills CLI (Vercel)
npx skills add ruvnet/ruflo --skill daa-agent -a claude-code2Manually: put the files in place
~/.claude/skills/daa-agent/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/daa-agent/
1Skills CLI (Vercel)
npx skills add ruvnet/ruflo --skill daa-agent -a codex2Manually: put the files in place
.agents/skills/daa-agent/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add ruvnet/ruflo --skill daa-agent -a cursor2Manually: put the files in place
.cursor/skills/daa-agent/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install ruvnet/ruflo daa-agent2Manually: put the files in place
.github/skills/daa-agent/SKILL.md1Skills CLI (Vercel)
npx skills add ruvnet/ruflo --skill daa-agent -a gemini-cli2Platform CLI
gemini skills install https://github.com/ruvnet/ruflo --path plugins/ruflo-daa/skills/daa-agentManually: put the files in place
.devin/skills/daa-agent/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add ruvnet/ruflo --skill daa-agent -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about daa-agent
How do I install daa-agent in Claude Code?
Run npx skills add ruvnet/ruflo --skill daa-agent -a claude-code. The skill then sits at ~/.claude/skills/daa-agent/SKILL.md.
What does daa-agent do?
It creates and adapts Dynamic Agentic Architecture agents that learn from interactions, change behavior based on metrics, and share knowledge with other agents.
How is a DAA agent different from a static one?
A static agent holds a fixed configuration and keeps its knowledge to itself. A DAA agent adapts over time, learns continuously and propagates what it learned to other agents.
Source text
SKILL.md · MIT
--- name: daa-agent description: Create and adapt Dynamic Agentic Architecture agents that learn and evolve argument-hint: "<create|adapt|status> [options]" allowed-tools: mcp__plugin_ruflo-core_ruflo__daa_agent_create mcp__plugin_ruflo-core_ruflo__daa_agent_adapt mcp__plugin_ruflo-core_ruflo__daa_learning_status mcp__plugin_ruflo-core_ruflo__daa_performance_metrics mcp__plugin_ruflo-core_ruflo__daa_knowledge_share Bash --- # DAA Agent Create agents with Dynamic Agentic Architecture that adapt and learn over time. ## When to use When you need agents that go beyond static configurations — agents that adapt their behavior based on performance metrics, learn from interactions, and share knowledge with other agents. ## Steps 1. **Create agent** — call `mcp__plugin_ruflo-core_ruflo__daa_agent_create` with initial configuration and learning parameters 2. **Monitor learning** — call `mcp__plugin_ruflo-core_ruflo__daa_learning_status` to see adaptation progress 3. **Check performance** — call `mcp__plugin_ruflo-core_ruflo__daa_performance_metrics` for efficiency and accuracy metrics 4. **Adapt** — call `mcp__plugin_ruflo-core_ruflo__daa_agent_adapt` to trigger manual adaptation based on feedback 5. **Share knowledge** — call `mcp__plugin_ruflo-core_ruflo__daa_knowledge_share` to propagate learnings to other agents ## DAA vs static agents | Aspect | Static Agent | DAA Agent | |--------|-------------|-----------| | Behavior | Fixed configuration | Adapts over time | | Learning | None | Continuous from interactions | | Knowledge | Isolated | Shared across agents | | Performance | Constant | Improves with use |
© ruvnet · MIT · ruvnet/ruflo
