Agent Entry and Bootstrap
Resolve an adopted project to its governing workspace and give every supported agent host the same bounded, verified starting point.
An agent should not begin with an unbounded repository scan. Workspai publishes a portable project entry, resolves the governing workspace locally, and verifies a bounded read order before the model starts broad discovery.
Bootstrap from the project
npx workspai agent bootstrap --for-agent generic --strict --jsonThe receipt checks project membership, workspace identity, live-input fingerprints, evidence freshness, the active Goal binding, Skills, and the Graph retrieval route. It does not persist a machine-local absolute workspace path in portable source.
Use project agent-entry when you need to inspect or validate the entry adapters
without compiling the complete bootstrap receipt:
npx workspai project agent-entry --jsonCanonical read order
.workspai/agent-entry.v1.json- the strict agent bootstrap receipt
.workspai/reports/INDEX.json- the active Goal handoff and relevant Skills
- scoped Graph queries and source inspection
The entry is a locator and integrity boundary. The CLI remains the authority for workspace identity, freshness, verification, and repair closure.
Failure behavior
A missing, stale, incompatible, or ambiguous entry fails closed with a structured reason. Consumers should refresh or repair the named producer, then bootstrap again. They should not silently select a different workspace or widen project scope.
See the contract-derived command pages for
agent bootstrap,
project agent-entry, and
workspace context.
Agent Grounding
How Workspai aligns Copilot, Cursor, Claude, Codex, MCP clients, and other agents around the same evidence.
Agent Skills and Grounded Context
Learn how Workspai gives coding agents shared instructions, focused graph evidence, and current workspace context without turning prompts into a second source of truth.