llm-ingestion-okf/docs/plan
Kjell Tore Guttormsen ff9f2f82ca docs(roadmap): phase 5 — MCP as a way to populate a bundle
Committed by the operator; the design is open and the scope is not. Recorded
in the roadmap with a scoping plan rather than in the v0.2 alignment doc, which
is about OKF v0.2 conformance and is the wrong home.

The plan settles nothing on purpose. It names the fork everything turns on —
whether we are the MCP server (an agent calls our doors, Door B's territory,
ours to design) or an MCP client (a manifest source type, which is normative in
commons' ingest-spec and therefore starts as a request to them, not as code
here). The two share a protocol and nothing else, including their owner.

Constraints written down because each has already refused something: the
one-runtime-dependency rule (an MCP SDK would be the second — stdlib JSON-RPC
or an optional extra), no model calls in the run path (the model may CALL us,
never the reverse), the guard persist gate (MCP is transport, not a bypass),
the network opt-in, and the required ingested_at.
2026-07-27 20:12:58 +02:00
..
execution-order.md docs(okf-v0.2): requirement 1 re-sized -- the emitter does not change at all 2026-07-26 19:57:29 +02:00
mcp-bundle-population.md docs(roadmap): phase 5 — MCP as a way to populate a bundle 2026-07-27 20:12:58 +02:00
okf-v0.2-alignment.md docs(okf-v0.2): U7 resolved — the 'not:' family is an example, not a field 2026-07-27 16:32:40 +02:00
phase-1-door-a.md docs(plan): add detailed phase 1-4 implementation plans 2026-07-16 10:58:08 +02:00
phase-2-doors-b-c.md feat(guard): wire Doors B and C to the real guard (Phase 2 step 4) 2026-07-25 07:32:45 +02:00
phase-3-configurable-contract.md docs(okf-v0.2): the alignment track lands before the contract freezes 2026-07-26 11:42:37 +02:00
phase-4-node-half.md docs(phase-4): record linkedin-studio's two named requirements before the freeze 2026-07-25 20:34:54 +02:00