Connect and structure what Capture and Builder produce.
ContextCurator is the organization surface that links ContextCapture and ContextBuilder outputs into organized, governed context objects. Kanban-style management, readiness scoring, and agent work packet generation turn raw context into execution-ready intelligence.
Captured knowledge without organization is just noise.
Ingesting documents and collecting employee contributions produces raw material — not governed context. Without an organization layer, captured knowledge remains disconnected fragments. Documents do not link to the employee insights that explain their exceptions. Contributions do not connect to the source material that validates them. Nothing tracks whether context is ready for agent consumption or still needs human review.
ContextCurator solves this by providing a Kanban-style workspace where teams connect, structure, and govern the outputs of ContextCapture and ContextBuilder. Readiness scoring ensures only sufficiently structured context reaches agents. Work packet generation turns governed context objects into execution-ready artifacts.
From raw fragments to governed context objects.
Outputs from ContextCapture (parsed documents, videos, audio) and ContextBuilder (employee contributions, Q&A, demand signals) flow into the curation workspace.
Kanban-style context management lets teams group, prioritize, and track context objects through governance stages. Visual boards show what is ready, what needs review, and what is blocked.
Connect Capture outputs to Builder contributions. Link source documents to the employee insights, edge cases, and workflow friction that explain how they are actually used.
Readiness scoring evaluates whether context objects have sufficient structure, connections, and verification for safe agent consumption. Incomplete objects cannot proceed.
Agent work packets are generated from governed context objects — packaging intent, constraints, source context, and acceptance criteria into structured execution artifacts.
Everything between raw context and agent-ready execution.
Every governed context object can emit a structured execution artifact.
The agent work packet wraps human intent in everything an agent needs to act safely. Source context, constraints, acceptance criteria, and a readiness score — all versioned, all traceable.
// agent_work_packet.json
{
"packet_id": "awp_2c91",
"issue_id": "PROJ-482",
"intent": {
"goal": "Add rate limit to /api/contact endpoint",
"actor": "engineer:michael",
"motivation": "Prevent abuse of strategy session form"
},
"constraints": [
"must not modify auth middleware",
"backwards compatible with existing clients",
"latency budget: < 5ms p99"
],
"source_context": [
"atom_8f3a", // rate limiter design decision
"atom_a02d" // API style guide
],
"acceptance_criteria": [
"returns 429 with Retry-After header",
"unit tests cover burst + sustained",
"OTel span emitted on throttle event"
],
"readiness_score": 0.82,
"readiness_gate": 0.75,
"agent_authorized": true,
"trace_id": "7c4f...a912"
}