Project Management

Seven tools for creating projects, loading project data, saving artifacts, triggering renders, and generating reviews.

---

init_project

Create a new project with the standard directory structure. Sets up folders for briefs, scenes, manifests, renders, and reviews.

Returns: Project path, slug, and created directory structure.

Try asking your AI:

"Initialize a new project called product-launch-q2"

---

list_projects

List all projects in the workspace.

No parameters required.

Returns: Array of project slugs with names, descriptions, and last modified timestamps.

Try asking your AI:

"What projects do I have?"

---

get_project

Load a specific project by slug, including its current state and artifacts.

Returns: Project metadata, current brief, latest manifest, render history, and review status.

Try asking your AI:

"Load the fintech-demo project"

---

get_project_context

Get enriched project context for AI direction. Aggregates brand, brief, scenes, and previous scores into a single context object used by the autonomous direction pipeline.

Returns: Aggregated context with brand package, brief, scene inventory, score history, and revision log.

Try asking your AI:

"Get the full project context for the fintech-demo project"

---

save_project_artifact

Save an artifact to a project. Artifacts are typed: brief, manifest, render, or review. Each save creates a versioned entry.

Returns: Saved artifact path, version number, and timestamp.

Try asking your AI:

"Save this manifest as an artifact to the product-launch project"

---

render_project

Trigger a render for a project. Runs the full render pipeline: resolve targets, capture/render, assemble, and encode for delivery.

Returns: Render job status, output file path, and encoding details.

Try asking your AI:

"Render the product-launch project for web-hero delivery"

---

review_project

Generate a comprehensive review of a project. Scores the latest manifest, audits annotations, checks brand compliance, and produces a summary report.

Returns: Review report with overall score, per-dimension scores, annotation quality, brand compliance status, and improvement recommendations.

Try asking your AI:

"Generate a review for the fintech-demo project"