Generate comprehensive technical documentation by synthesizing all accumulated insights from the neural memory knowledge graph. Also use to save a new insight with neural_add_insight.
Scanned 5/27/2026
Install via CLI
openskills install Yakoub-ai/neural-memory---
name: neural-insight
description: Generate comprehensive technical documentation by synthesizing all accumulated insights from the neural memory knowledge graph. Also use to save a new insight with neural_add_insight.
---
# Neural Memory — Insight Bank
The insight bank accumulates technical knowledge about the project — implementation choices, architecture decisions, and patterns discovered during development. Use `/neural-insight` to synthesize everything into structured documentation.
## Generate full technical documentation
Use the `neural-doc-writer` agent (bundled with this plugin) to synthesize all accumulated insights:
```
Use the neural-doc-writer agent to generate documentation
```
Or call the MCP tool directly:
```json
Tool: neural_generate_docs
{
"project_root": "."
}
```
Output is returned as markdown **and** written to `.neural-memory/technical-docs.md`.
## Bundled agents
This plugin includes three agents installed alongside the MCP server and skills:
| Agent | Purpose |
|-------|---------|
| `neural-memory:neural-explorer` | Codebase exploration via semantic graph search |
| `neural-memory:neural-insight-collector` | Captures insights from conversations into the bank |
| `neural-memory:neural-doc-writer` | Synthesizes all insights into technical documentation |
These agents have the neural-memory MCP tools pre-configured — no extra setup needed.
## Save an insight
```json
Tool: neural_add_insight
{
"content": "The bump script atomically updates 4 files to keep versions in sync...",
"topic": "versioning",
"related_files": ["scripts/bump_version.py"]
}
```
Insights are deduplicated — re-saving the same insight updates it rather than creating a duplicate.
## Browse insights
```json
Tool: neural_list_insights
{
"topic": "storage"
}
```
Omit `topic` to list all insights grouped by topic.
## Parameters — neural_add_insight
| Parameter | Required | Type | Description |
|-----------|----------|------|-------------|
| `content` | Yes | str | The insight text (min 10 chars) |
| `topic` | Yes | str | Topic area, e.g. `storage`, `hooks`, `embeddings`, `cli` |
| `related_files` | No | list[str] | File paths to link via RELATES_TO edges |
| `project_root` | No | str | Default: `"."` |
## Parameters — neural_generate_docs / neural_list_insights
| Parameter | Required | Type | Description |
|-----------|----------|------|-------------|
| `project_root` | No | str | Default: `"."` |
| `topic` | No | str | (`list_insights` only) Filter by topic |
## Notes
- Insights accumulate without a lifecycle — no status or archiving
- The more insights saved during development sessions, the richer the generated documentation
- `session_context.md` includes an insight bank summary (topic counts) automatically
- Use `neural_query` to search insights alongside code nodes

No comments yet. Be the first to comment!