Back to index

The Project Loop

Your collaborative workflow — from editorial brief to production code, with full context shared between you and your AI assistant.

01 — The Daily Cycle

Agiflow provides a systematic approach for every phase of your project. Your AI assistant uses these shared skills to manage projects, refine work, and report progress.

Plan
Refine
Groom
Execute
Track

02 — Core Operations

Planning

Editorial phase

Break high-level vision and briefs into concrete, manageable tasks.

Key actions

  • Create focused task lists
  • Define clear success criteria
  • Group tasks into feature units
  • Establish priorities

Refinement

Technical context

Add detail and architectural guidance to tasks before your assistant begins.

Key actions

  • Add implementation details
  • Document project conventions
  • Link related requirements
  • Assign to specific assistants

Grooming

Prioritization

Maintain a clean and organized queue for your team.

Key actions

  • Reorder tasks by importance
  • Manage feature boundaries
  • Identify blocking items
  • Close obsolete requests

Review

Verification

Check completed work, provide feedback, and iterate on results.

Key actions

  • Audit assistant output
  • Provide editorial feedback
  • Track through status stages
  • Review detailed progress logs

03 — Execution Pipeline

Implementation

When it's time to build, your AI assistant follows a structured implementation path to ensure quality and consistency.

1

Load context

The assistant reads the task requirements, success criteria, and related architectural context.

2

Observe patterns

Before making changes, it identifies established project conventions and design rules.

3

Build

It implements the solution, strictly following the editorial vision and technical patterns.

4

Validate

After building, it verifies the work against project-specific rules and success criteria.

5

Update progress

The assistant reports the results, updates task status, and documents the changes.

04 — Shared Knowledge

Progress Tracking

Work units maintain detailed metadata that both you and your AI assistant can access. This ensures perfect continuity — everyone knows what has changed and what remains.

Example progress log
{ "progress": { "plan": "Database → Service → UI", "status": { "completed": 2, "total": 6 }, "changes": [ "backend/schema.sql", "packages/api/client.ts" ], "verification": { "passed": 12, "quality": "high" } } }

05 // Related

Support

Need help getting a board connected to your assistant?

Email support

Feedback

Missing a guide for your team’s project setup?

Open an issue

Community

Share how your team plans campaigns, deals, and client work.

Join Discord