| name | mofa-agent-memory-deck |
| description | Visualizes an agent's internal thought process, reasoning, and memory observations into a professional slide presentation. |
MoFA Agent Memory Deck
This skill ingests an agent's memory logs—thoughts, decisions, and system observations—and synthesizes them into an engaging PowerPoint (.pptx) presentation. It bridges backend memory systems (like mofa-memory) with human-readable visual reporting, creating transparent, auditable slides of an agent's execution history.
Trigger Phrases
Activate this skill when user says:
- "Create a memory presentation"
- "Audit the agent's thought process"
- "Generate a slide deck from the agent trace"
- "Show me what the agent did visually"
- "Memory to Slides"
Usage (CLI Examples)
mofa slides \
--slide-dir ./mofa-slides \
--style memory-review \
--input agent_memory_slides.json \
--out memory_audit.pptx
Examples
The skill expects a raw JSON payload describing the agent's memory execution trace over a specific session. You can find a sample payload in examples/mock_agent_trace.json:
This skill conceptually starts from a raw JSON payload describing the agent's memory execution trace over a specific session. You can find a sample trace payload in examples/mock_agent_trace.json:
{
"agent_id": "Research-Agent-02",
"task": "Analyze distributed consensus algorithms for agents",
"traces": [
{ "timestamp": "2026-03-16T10:00:00Z", "type": "Observation", "content": "Found 12 research papers on Raft networks." },
{ "timestamp": "2026-03-16T10:05:00Z", "type": "Decision", "content": "Filtered out papers older than 2024 to ensure relevancy." }
]
}
In practice, this trace-shaped JSON is first transformed (by separate tooling or an upstream skill) into a slides input of the form:
[
{ "prompt": "Summarize the overall task and agent objective for the deck intro slide." },
{ "prompt": "Explain the key observations the agent made about Raft networks." }
]
This slides JSON (for example, agent_memory_slides.json) is what you pass to the mofa slides CLI via --input. The memory-review style must be available under the mofa-slides/styles/ directory (for example, mofa-slides/styles/memory-review.toml) so that --style memory-review can be resolved. The style template acts as a prefix instructing the model on how to organize the slides (e.g., Executive Summary, Key Decisions, Future Context) before passing the results to the standard mofa slides backend pptxgenjs/Rust generator.
Features
- Observatory Analytics: Brings "glass-box" visibility to MoFA swarms by turning raw memory JSON into client-ready presentations.
- Auditable Context: Perfect for analyzing how an agent reached a specific decision during a complex workflow.
- Synergy: Serves as the visual frontend companion to backend memory storage paradigms like SQLite vector databases.