# aria-knowledge

aria-knowledge builds a local Markdown memory base for Claude Code, preserving architecture choices and debugging context across sessions under human review.

- plugin · `type=plugin` · Claude Code · ★ 19 repo stars
- Author: mikeprasad
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/mikeprasad/aria-knowledge/tree/HEAD/plugin
- Updated: 2026-10-22
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/plugins/aria-knowledge

## What it does

This plugin stops agents from starting from zero every time a session resets. When context compaction clears transcripts, key debugging insights and design agreements vanish. The tool stores this operational memory locally inside `~/Projects/knowledge/` as standard Markdown files.

Knowledge moves through defined capture and structuring steps. The `/extract` command scans conversations for fresh findings and removes duplicates. External files and URLs enter through `/intake`, while `/distill` and `/codemap` connect task tickets directly to repository paths. To carry the active state into the next run, use `/handoff`.

Captured items never become verified rules on their own. Every entry waits in a staging backlog first. You must deliberately review and promote records via `/audit-knowledge` before the assistant can treat them as trusted project facts.

## Install

### Claude Code (native platform)

**Plugin marketplace**
In the agent chat (not a shell):
```text
/plugin marketplace add anthropics/claude-plugins-community
```
In the agent chat (not a shell):
```text
/plugin install aria-knowledge@claude-community
```

## When to use it
- retain decisions and debugging discoveries across chat resets
- prevent reasoning continuity breaks caused by context compaction
- import external docs and links into a structured staging queue
- generate repository-grounded task specifications using `/distill`
- transfer full session context to future agent runs with `/handoff`

## Example prompts
- Run `/extract` to harvest recent decisions and lessons from the dialogue
- Use `/distill` to create a grounded task spec citing actual repository files
- Execute `/audit-knowledge` to inspect backlog entries and promote verified knowledge

## Similar
- [compile-knowledge](https://affpapa.org/en/nevibecode/skills/compile-knowledge.md) — skill · 3300 installs
- [knowledge-base](https://affpapa.org/en/nevibecode/skills/knowledge-base.md) — skill · 875 installs
- [cabinet-ai-knowledge-base](https://affpapa.org/en/nevibecode/skills/cabinet-ai-knowledge-base.md) — skill · 615 installs
- [wiki-knowledge-complier](https://affpapa.org/en/nevibecode/plugins/wiki-knowledge-complier.md) — plugin · ★ 12 repo stars
- [knowledge-base-builder](https://affpapa.org/en/nevibecode/skills/mario-aldayuz-littlebird-knowledge-base-builder.md) — skill · ★ 96 repo stars
- [distill-memory](https://affpapa.org/en/nevibecode/skills/distill-memory.md) — skill · 743 installs

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
