pruuf

AI agent portfolio/outcome tracker — log proof of what AI has accomplished.

What It Does

pruuf creates a portfolio of AI agent accomplishments. When users celebrate, show positive sentiment, or explicitly ask to log outcomes, pruuf records:

• What was accomplished
• Evidence/proof (screenshots, links, quotes)
• Impact metrics (time saved, value created)
• The trigger that prompted the log

Over time, this builds a database showing the outcomes an AI agent has created.

Quick Start

npm install @cubicrew/pruuf
pruuf init           # creates .pruuf/ directory
pruuf serve          # start REST API + dashboard on :3457

Or use as an MCP server:

pruuf mcp            # start MCP server (stdio)

CLI Commands

| Command | Description | |---------|-------------| | pruuf init | Initialize a .pruuf/ data directory | | pruuf mcp | Start MCP server (stdio) | | pruuf serve [-p PORT] | Start REST API server + dashboard | | pruuf log <title> | Log an outcome with full details | | pruuf quick <title> | Quick log with just a title | | pruuf list | List outcomes with filters | | pruuf get <id> | Get outcome by ID | | pruuf update <id> | Update an outcome | | pruuf delete <id> | Delete an outcome | | pruuf summary | Show portfolio stats | | pruuf report | Generate portfolio report (HTML/Markdown) | | pruuf categories | List outcome categories | | pruuf evidence <id> | Add evidence to an outcome | | pruuf link <outcomeId> <clawckId> | Link outcome to clawck time entry | | pruuf portfolio export | Export portfolio as static HTML | | pruuf sync | Trigger sync from remote sources |

Global options: --json, -d, --dir <path>

MCP Tools

| Tool | Description | |------|-------------| | log_outcome | Full outcome logging with evidence and impact | | quick_log | Fast logging with minimal info | | log_celebration | Log from user celebration/positive sentiment | | get_outcome | Get specific outcome by ID | | query_portfolio | Filter outcomes by category, project, date | | get_portfolio | Get full portfolio | | get_summary | Get stats (by category, project, totals) | | add_evidence | Add proof to existing outcome | | update_outcome | Update outcome details | | delete_outcome | Remove outcome | | list_categories | List outcome categories | | generate_report | Generate portfolio report (HTML/Markdown) | | link_clawck_entry | Link an outcome to a clawck time entry |

REST API

When running pruuf serve, endpoints are available at http://localhost:3457/api:

• GET /api/health — Health check
• GET /api/stats — Portfolio statistics
• POST /api/outcomes — Create outcome
• GET /api/outcomes — List outcomes (query params: category, project, client, tag, from, to, limit)
• GET /api/outcomes/:id — Get outcome
• PATCH /api/outcomes/:id — Update outcome
• DELETE /api/outcomes/:id — Delete outcome
• POST /api/outcomes/:id/evidence — Add evidence
• GET /api/agents — List agents
• GET /api/portfolio/:agentId — Get portfolio
• POST /api/reports/generate — Generate report

Categories

• deliverable — Completed work product
• automation — Automated a process
• research — Research/analysis completed
• communication — Successful communication
• problem-solved — Fixed an issue
• creative — Creative work (design, writing)
• integration — Connected systems
• optimization — Improved performance
• other — Uncategorized

Environment Variables

• PRUUF_DIR — Where to store portfolio data (default: .pruuf)
• PRUUF_AGENT_ID — Agent identifier (default: default-agent)

License

MIT — CubiCrew 2026