Tianyi Cui 00d7646558 fix review findings: make the prune-seam RFC + index match the persistence-only shipped scope ...

The reviewer caught two pieces of both-seams drift left over after the bash
get()/list() removal was reverted to a persistence-only change.

- docs/rfc/README.md: rename the index row from "persistence and bash seams" to
  "Prune dead methods from the persistence seam" so it matches the RFC title and
  the actually-shipped scope (verify-rfc-classification only checks the path is
  indexed, so this prose slipped the gate).
- The implemented RFC body still read like the original both-seams proposal
  (the "Two capability seams" framing, a `### BashExecutor.get()/.list()` problem
  section, a bash removal bullet in the Proposal, and current-source links that
  imply bash get/list were removed). Rewrite the body into the durable
  decision-record form: Problem/Proposal/criteria/risks now describe only the
  persistence has()/delete() removal that shipped, and the bash reasoning (why
  get()/list() earn their keep — a ~35-line test-harness migration cost makes the
  test consumer a real consumer) is folded into the top decision note as
  "considered and deliberately kept", not as a shipped change. Drop the stale
  bash source-line refs; keep the persistence consumer links pointing at current
  code (agent-loop load, ACP session/list).

2026-06-21 11:05:56 +08:00

..

2026-06-19-drop-mutable-session-summary.md

docs(rfc): classify RFCs by kind via path-encoded subdirectories

2026-06-20 22:29:45 +08:00

2026-06-20-drop-unconsumed-llm-adapter-change-event.md

fix review findings: stale service prose + catalog cleanup

2026-06-21 01:41:02 +08:00

2026-06-20-drop-unconsumed-llm-assembled-surfaces.md

simplify(llm): drop unconsumed adapter-change event and assembled call surfaces

2026-06-21 01:27:41 +08:00

2026-06-20-prune-dead-seam-methods.md

fix review findings: make the prune-seam RFC + index match the persistence-only shipped scope

2026-06-21 11:05:56 +08:00