memory-transform
Exports agent memory into human-readable visual formats such as Mermaid diagrams, static dashboards, and Obsidian vaults.
Install
mkdir -p .claude/skills/memory-transform && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/15942" && unzip -o skill.zip -d .claude/skills/memory-transform && rm skill.zipInstalls to .claude/skills/memory-transform
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Agent memory transformation skill. Converts the indexed agent memory into navigable representations including Obsidian vaults with backlinks, Mermaid dependency and status graphs, Graphviz DOT files, and static dashboards. Use when the user wants to visualize requirements, browse memory in Obsidian, generate a status graph, export a dependency diagram, produce a memory report, or navigate agent memory in any graphical format. Never mutates the canonical memory under docs/agent-memory/ — all output goes to a separate path.Key capabilities
- →Convert indexed agent memory into Mermaid dependency and status graphs
- →Generate Graphviz DOT files for richer layouts and cluster grouping
- →Export memory into Obsidian vaults with backlinks for browsable knowledge bases
- →Produce human-readable Markdown status reports from memory
- →Generate static HTML dashboards for standalone browser views
- →Read `docs/agent-memory/index.json` and folder-level `index.json` files for structure
How it works
The skill reads the indexed agent memory and uses various scripts to transform it into different navigable formats like Mermaid graphs, Obsidian vaults, or static dashboards, ensuring the original memory remains unchanged.
Inputs & outputs
When to use memory-transform
- →Visualizing agent memory
- →Generating status reports
- →Creating Obsidian vaults from data
- →Exporting dependency diagrams
About this skill
Memory Transform Skill
Overview
Converts indexed agent memory (docs/agent-memory/index.json) into navigable output formats. The canonical memory is never modified — all generated files go to a separate destination.
0) References and Scripts
Resolve <SKILL_DIR> from the path used to load this file (replace SKILL.md with empty).
| Need | File |
|---|---|
| Mermaid, DOT, and Obsidian conventions and output structure | <SKILL_DIR>/references/format-guide.md |
| Export memory to any supported format | <SKILL_DIR>/scripts/export-memory-graph.mjs |
| Write a Markdown status report | <SKILL_DIR>/scripts/memory-report.mjs |
| Export an Obsidian vault | <SKILL_DIR>/scripts/obsidian-vault.mjs |
| Generate a static dashboard | <SKILL_DIR>/scripts/memory-dashboard.mjs |
Read the format guide before running an export — it explains the exact output structure, node conventions, and linking patterns for each format.
1) Choose a Format
| Format | Command | Best for |
|---|---|---|
| Mermaid | node <SKILL_DIR>/scripts/export-memory-graph.mjs --format=mermaid | Status overviews, quick diagrams |
| DOT (Graphviz) | node <SKILL_DIR>/scripts/export-memory-graph.mjs --format=dot --out=<path> | Richer layouts, cluster grouping |
| Obsidian vault | node <SKILL_DIR>/scripts/obsidian-vault.mjs --out=<dir> | Browsable knowledge base with backlinks |
| Markdown report | node <SKILL_DIR>/scripts/memory-report.mjs --out=<path> | Human-readable status digest |
| HTML dashboard | node <SKILL_DIR>/scripts/memory-dashboard.mjs --out=<path> | Standalone browser view |
Full command options:
# Mermaid — print to stdout
node <SKILL_DIR>/scripts/export-memory-graph.mjs --format=mermaid
# Mermaid — write to file
node <SKILL_DIR>/scripts/export-memory-graph.mjs --format=mermaid --out=docs/reports/memory-graph.md
# DOT — write to file and render with Graphviz
node <SKILL_DIR>/scripts/export-memory-graph.mjs --format=dot --out=docs/reports/memory.dot
dot -Tsvg docs/reports/memory.dot -o docs/reports/memory.svg
# Obsidian vault
node <SKILL_DIR>/scripts/obsidian-vault.mjs --out=docs/reports/obsidian-vault/
# Markdown report
node <SKILL_DIR>/scripts/memory-report.mjs --out=docs/reports/memory-report.md
# Static dashboard
node <SKILL_DIR>/scripts/memory-dashboard.mjs --out=docs/reports/memory-dashboard.html
# Use a different index file
node <SKILL_DIR>/scripts/export-memory-graph.mjs --format=mermaid --index=<path-to-index.json>
2) After Export
- Tell the user exactly what files were generated and where
- If the format is Mermaid: render inline if the output channel supports it, otherwise provide the file path
- If the format is DOT: tell the user how to render it (
dot -Tpng, or paste into graphviz.online) - If the format is Obsidian: tell the user to open
<output-dir>as an Obsidian vault
3) Reading Memory for Reports
When the user asks for a narrative report or dashboard summary (not a graph):
- Read
docs/agent-memory/index.jsonfor the top-level structure - Read folder-level
docs/agent-memory/**/index.jsonfiles for per-folder artifact lists - Read
docs/agent-memory/08-progress-index.mdfor the human-readable status table - Synthesize — do not copy raw JSON into the report; transform it into readable prose or tables
The canonical memory is the source; never invent status, dates, or links.
4) Canonical Memory Rules
- Never write to
docs/agent-memory/as part of a transform - Output only to
docs/reports/, a user-specified path, or an explicit temp location - List all generated files in the skill output so the user knows what was created
When not to use it
- →When mutating the canonical memory under `docs/agent-memory/`
- →When the user does not want to visualize or navigate agent memory
- →When the user requests a narrative report that copies raw JSON
Limitations
- →Never mutates the canonical memory under `docs/agent-memory/`
- →All output goes to a separate path, typically `docs/reports/`
- →Requires `node` for script execution
How it compares
This workflow provides multiple structured and visual representations of agent memory without altering the source data, offering diverse ways to interact with and understand the memory compared to directly reading raw index files.
Compared to similar skills
memory-transform side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| memory-transform (this skill) | 0 | 3mo | Review | Intermediate |
| readme-metrics-dashboard | 0 | 4mo | No flags | Beginner |
| meeting-minutes | 41 | 6mo | No flags | Beginner |
| notion-knowledge-capture | 10 | 9mo | No flags | Intermediate |
Try saying
Example prompts that trigger this skill in your AI assistant.
You might also like
readme-metrics-dashboard
harrybin
Refresh repository metrics, render a full chart dashboard (line, bar, area, pie, scatter, composed), and update README with a reproducible example.
meeting-minutes
github
Generate concise, actionable meeting minutes for internal meetings. Includes metadata, attendees, agenda, decisions, action items (owner + due date), and follow-up steps.
notion-knowledge-capture
makenotion
Transforms conversations and discussions into structured documentation pages in Notion. Captures insights, decisions, and knowledge from chat context, formats appropriately, and saves to wikis or databases with proper organization and linking for easy discovery.
super-save
supermemoryai
Save important project knowledge to memory. Use when user wants to preserve architectural decisions, significant bug fixes, design patterns, or important implementation details for team reference.
knowledge-base
incidentfox
Search runbooks, documentation, and knowledge base articles from Confluence. Use when looking for incident response procedures, service documentation, post-mortems, or troubleshooting guides.
decision-logger
alirezarezvani
Two-layer memory architecture for board meeting decisions. Manages raw transcripts (Layer 1) and approved decisions (Layer 2). Use when logging decisions after a board meeting, reviewing past decisions with /cs:decisions, or checking overdue action items with /cs:review. Invoked automatically by the board-meeting skill after Phase 5 founder approval.