Skip to main content

File Explorer

Effective version

v10.0.179+ (standing instructions in instructions.md).

Each Agent's configuration, memory, and skills are stored as files in AgentFS (Agent File System). The file explorer lets you intuitively browse and manage these files.

AgentFS File Structure Overview​

Each Agent's files are stored in the ~/.desirecore/agents/<agent_id>/ directory:

<agent_id>/
+-- agent.json        # Entry configuration
+-- persona.md        # Persona
+-- principles.md     # Principles
+-- instructions.md   # Optional stable responsibilities, strategies, and delivery standards
+-- memory/           # Memory files
|   +-- *.md          # One file per memory
+-- skills/           # Skills directory
|   +-- <skill_id>/   # One directory per skill
|       +-- SKILL.md  # Skill documentation
+-- tools/            # Tool registration
+-- workflows/        # Workflow definitions
+-- schedules/        # Scheduled task definitions
+-- heartbeat/        # Heartbeat configuration
+-- resources/        # Resource files
+-- assets/           # Static resources
+-- .git/             # Git version management
Why Use File System?

DesireCore's design philosophy is "everything is files." Files are more transparent than databases - you can directly view, edit, back up, and version manage them. Every change to an Agent is recorded in Git history.

Browsing Agent Files​

Opening the File Explorer​

  1. Enter the Agent details page
  2. Click the "Files" tab
  3. The file explorer displays all the Agent's files in a tree structure

File Preview​

Click any file to preview its content on the right:

  • Markdown files (.md): Rendered rich text preview
  • JSON files (.json): Syntax-highlighted JSON view
  • YAML files (.yaml): Syntax-highlighted YAML view

File Editing​

For Markdown and configuration files, you can edit directly in the preview interface. After editing and saving, a Git commit is automatically generated and recorded in version history.

Edit Configuration Files with Caution

Directly editing core configuration files like agent.json may affect the Agent's normal operation. If you're unsure about the meaning of a field, it's recommended to modify through the interface or conversation.

Edit Standing Instructions​

Open an existing instructions.md in the file tree and edit it like other Markdown files, or select the agent's Standing Instructions in Prompt Center. If the file does not exist yet, ask the agent through conversation to save its stable responsibilities, then edit the body.

This file has no L0/L1/L2 loading tiers. Saved changes are read when the next input, explicit continuation, or restoration is accepted; the current execution turn keeps its original snapshot. Put detailed SOPs in skills and personal preferences in the user scope. See File Format Reference for full-body loading, budget, and update behavior.

Uploading and Managing Resource Files​

You can upload reference documents, templates, data files, etc. to the Agent's resources/ directory for the Agent to reference during work.

Uploading Files​

  1. Navigate to the resources/ directory in the file explorer
  2. Click the "Upload" button
  3. Select local files
  4. Files are automatically committed after upload

Managing Files​

  • Rename: Right-click the file, select "Rename"
  • Delete: Right-click the file, select "Delete" (recorded in version history, recoverable)
  • Move: Drag and drop files to the target directory

File Type Reference​

FileFormatDescription
agent.jsonJSONEntry configuration: name, version, description, default model, environment injection, heartbeat switch, permissions, and repository settings
persona.mdMarkdownPersona: tone, style, response strategy
principles.mdMarkdownPrinciples: rules, forbidden zones, priorities
instructions.mdOrdinary MarkdownOptional agent-specific stable responsibilities, default work strategies, and delivery standards; loaded in full when allowed
memory/_policy.jsonJSONMemory compression and retention policy
memory/*.mdMarkdownMemory entries: learned knowledge and preferences
skills/*/SKILL.mdMarkdownSkill documentation: skill description and execution instructions
tools/*.jsonJSONTool registration: parameter definitions for callable tools
heartbeat/HEARTBEAT.mdMarkdownHeartbeat inspection content: what the agent monitors and how it reports
workflows/*.yamlYAMLVisual workflow DSL
schedules/*.jsonJSONScheduled task definitions: trigger rule, target prompt, status, and lifecycle controls
resources/*AnyReference resources: documents, templates, data files
Editing Configuration Directly

Use agent.json.llm for default models, agent.json.env for date/time injection and date rollover reminders, heartbeat settings or agent.json.heartbeat.enabled for the heartbeat switch, and memory/_policy.json for automatic memory compression.

Next Steps​