Anamnesis Cloud

A VS Code extension that builds, uploads, and visualizes Anamnesis knowledge graphs on Anamnesis Cloud. Use it for AI-assisted architecture, dependency analysis, and impact exploration across multi-language codebases.

What's new

• Credentials for AI — Store encrypted key/value credential sets in the sidebar and tag them to one or more Anamnesis projects. Cursor agents can list and decrypt them locally via MCP.
• Project Prompts — Manage per-project prompts with Prompt Parameters, optional Skip verification via AI Model, and a View action that renders original, AI-generated, and parameter text as markdown.

Key features

• Create knowledge graphs — Right-click a workspace folder → Anamnesis: Create Knowledge Graph
• Projects tree — Browse graphs stored on the Anamnesis Cloud Server
• Project Prompts — Add, view, edit, and delete prompts for a selected project (parameters + optional skip-AI save)
• Credentials for AI — Encrypted named credential sets, multi-project tags, and local MCP tools for agents
• Dual viewer
  • Table view (default): fast render and text filtering for large graphs
  • Graph view: Cytoscape.js force-directed, circle, grid, concentric, or preset layouts
• Node inspection — Label, kind, community, source file, line, and neighbors
• Click to source — Open the original file at the correct line from the graph
• AI tools — Install MCP & Skill (Settings) or Anamnesis: Enable AI Tools (MCP + Skill) registers graph MCP plus a localhost anamnesis-credentials MCP server and an agent skill. Paths follow the host IDE (Cursor vs VS Code).
• Settings panel — Server URL, credentials, default tag, Test Connection, Save, and Install MCP & Skill

Supported project types

Anamnesis scans a workspace folder and produces one unified graph per project. Multiple extractors run in parallel; edges link code, config, build, and documentation layers together.

Application & backend code

Example: A Spring Boot repo gets Java class/method graphs and Maven module + dependency graphs in the same project graph.

Adobe Experience Manager (AEM)

Example: An AEM project like adobexp produces a graph connecting header.html → HeaderModel → _cq_dialog fields → ./headerTitle → Maven modules — enabling AI to find components, dialogs, and models without blind grep.

Infrastructure & DevOps

Apache and NGINX configs are auto-detected by path (sites-available, nginx, upstreams, etc.) and directive syntax.

Architecture documentation

Extractors reference

When you run Anamnesis: Create Knowledge Graph, these extractors are applied automatically:

Auto-detection notes

• Maven — Detected by pom.xml filename or Maven <project> / modelVersion in .xml files (not generic AEM .content.xml).
• HTL — Detected when the path is under apps/.../components/ or the file contains data-sly-* directives (takes precedence over plain HTML extraction).
• AEM .content.xml — Detected for JCR content files under jcr_root/apps/ (component definition, _cq_dialog, clientlibs, edit config).
• Plain HTML — Used only when the file is not classified as HTL.

Recommended scan settings

For Java/Maven/AEM repos, add build output folders to anamnesis.excludeGlobs to avoid duplicate nodes:

"anamnesis.excludeGlobs": ["target/", "dist/", "node_modules/"]

Default excludes already skip dist/, build/, out/, .git/, and node_modules/.

Project Prompts

The Project Prompts view in the Anamnesis activity bar lists prompts stored for each knowledge-graph project.

• Open a project row (or use Anamnesis: View Prompts on a folder) to load that project's prompt table.
• Add Prompt / Edit dialog fields:
  • Title and Original Prompt (required)
  • Prompt Parameters — optional values for placeholders or variables in a generic prompt (for example brand, locale, or task constraints)
  • Skip verification via AI Model — when checked, the original prompt is stored as-is (ready) and the server does not generate an AI-improved prompt
• Actions on each row:
  • View — markdown preview of Original Prompt, AI Generated Prompt, and Prompt Parameters, each with Copy
  • Edit / Delete
• Without skip-AI, create/update still queues AI improvement of the original prompt on the Anamnesis server.

Credentials for AI

The Credentials for AI view stores named key/value sets for agents (database URLs, API tokens, environment secrets, and similar). Ciphertext is saved on Anamnesis Cloud; values are decrypted only inside the extension.

• Toolbar + opens Add Credentials; click a set to edit it.
• Each set has a name, one or more key/value pairs, and optional Project tags.
• Tag a set with multiple Anamnesis projects so the same credentials apply across graphs. Older single-tag sets still load.
• After Install MCP & Skill, the agent in the current IDE can call:
  • list_credential_sets — names and metadata (no secret values)
  • get_credentials(name) — decrypt a named set locally and return key/value pairs
• Prefer a set whose project tags include the current graph when more than one match exists.

Quick start

1. Install the extension and open Anamnesis in the activity bar.
2. Configure Anamnesis Settings (Server URL, Client Id, Secret Key from Anamnesis Cloud). Use Install MCP & Skill so the current IDE (Cursor or VS Code) gets the graph MCP server and Anamnesis skill.
3. Right-click a project folder in the Explorer → Anamnesis: Create Knowledge Graph.
4. Open the project from the Projects tree to explore the graph.
5. Use Project Prompts to add reusable prompts (with parameters, or skip AI generation).
6. Use Credentials for AI to store encrypted secrets tagged to one or more projects.
7. Optional: if you skipped the Settings button, run Anamnesis: Enable AI Tools (MCP + Skill) from the command palette. The same installer detects Cursor vs VS Code and writes the matching MCP config and skill files.

Configuration

Anamnesis Cloud

Graphs are stored on Anamnesis Cloud. Create an account at anamnesis.cloud, configure credentials in the extension, and upload graphs directly from VS Code.

License

MIT