DuckDB Schema
clnode uses DuckDB as its local database. Data is stored at data/clnode.duckdb.
Tables
projects
| Column | Type | Description |
| id | VARCHAR PK | Project identifier |
| name | VARCHAR | Project name |
| path | VARCHAR UNIQUE | Filesystem path |
| created_at | TIMESTAMP | Auto-set via now() |
sessions
| Column | Type | Description |
| id | VARCHAR PK | Session identifier |
| project_id | VARCHAR | FK to projects |
| started_at | TIMESTAMP | Auto-set via now() |
| ended_at | TIMESTAMP | Set on session end |
| status | VARCHAR | active / ended |
agents
| Column | Type | Description |
| id | VARCHAR PK | Agent identifier |
| session_id | VARCHAR | FK to sessions |
| agent_name | VARCHAR | Display name |
| agent_type | VARCHAR | Role type |
| parent_agent_id | VARCHAR | FK to agents (parent) |
| status | VARCHAR | active / completed |
| started_at | TIMESTAMP | Auto-set via now() |
| completed_at | TIMESTAMP | Set on stop |
| context_summary | TEXT | Extracted work summary |
| input_tokens | INTEGER | Token usage (input) |
| output_tokens | INTEGER | Token usage (output) |
context_entries
| Column | Type | Description |
| id | INTEGER PK | Auto-increment |
| session_id | VARCHAR | FK to sessions |
| agent_id | VARCHAR | FK to agents |
| entry_type | VARCHAR | decision / blocker / handoff / agent_summary / note |
| content | TEXT | Entry content |
| tags | VARCHAR[] | Tags for filtering |
| created_at | TIMESTAMP | Auto-set via now() |
file_changes
| Column | Type | Description |
| id | INTEGER PK | Auto-increment |
| session_id | VARCHAR | FK to sessions |
| agent_id | VARCHAR | FK to agents |
| file_path | VARCHAR | Changed file path |
| change_type | VARCHAR | Edit / Write |
| created_at | TIMESTAMP | Auto-set via now() |
tasks
| Column | Type | Description |
| id | INTEGER PK | Auto-increment |
| project_id | VARCHAR | FK to projects |
| title | VARCHAR | Task title |
| description | TEXT | Task details |
| status | VARCHAR | idea / planned / pending / in_progress / needs_review / completed |
| assigned_to | VARCHAR | Agent name |
| tags | VARCHAR[] | Tags |
| created_at | TIMESTAMP | Auto-set via now() |
| updated_at | TIMESTAMP | Auto-set via now() |
| Column | Type | Description |
| id | INTEGER PK | Auto-increment |
| task_id | INTEGER | FK to tasks |
| author | VARCHAR | Comment author |
| comment_type | VARCHAR | plan / review / note |
| content | TEXT | Comment content |
| created_at | TIMESTAMP | Auto-set via now() |
activity_log
| Column | Type | Description |
| id | INTEGER PK | Auto-increment |
| session_id | VARCHAR | FK to sessions |
| agent_id | VARCHAR | FK to agents |
| event_type | VARCHAR | Event type |
| details | JSON | Event details |
| created_at | TIMESTAMP | Auto-set via now() |
DuckDB Notes
- Always use
now() — not current_timestamp COUNT(*) returns BigInt — wrap with Number()VARCHAR[] arrays need literal construction, bind params don’t work- Database file:
data/clnode.duckdb