Pieces Agent Skills — Development skill for Claude Code

Pieces Agent Skills

Two advanced agentic skills for Pieces OS, focusing on deep research synthesis and MCP-based long-term memory interactions.

One should be distributed to people who wish to integrate agents with Pieces MCP.

One should be used in your deep research report generation which is what you can see failing me in my original screenshots.

This, when given to an agent, is how I produced better reports from your existing framework. Any of the models can handle this kind of synthesis so there is little advantage to throwing Opus level reasoning models here.

Skills Included

1. [Deep Research Synthesizer](deep-research-synthesizer/SKILL.md)

Validates and synthesizes outputs from multiple research sub-agents into a coherent, citation-backed Markdown report.

• Key Features: Deduplication, contradiction handling, and explicit debug visibility.
• Location: deep-research-synthesizer/

2. [Pieces MCP Playbook](pieces-mcp-playbook/SKILL.md)

A procedural guide for reliably querying and writing to Pieces Long-Term Memory (LTM) via the Pieces MCP server.

• Key Features: Stable query/write patterns, troubleshooting, and request-shaping examples.
• Location: pieces-mcp-playbook/

Getting Started

Each skill contains its own `SKILL.md` with specific instructions, scripts, and reference documentation.

1. Explore the deep-research-synthesizer for multi-agent synthesis.
2. Explore the pieces-mcp-playbook for Pieces LTM integration.

License

MIT