feat(acl): unify the retrieval scopes; mark shared rows in Knowledge browse
CI & Build / Python lint (push) Successful in 3s
CI & Build / Python tests (push) Failing after 31s
CI & Build / TypeScript typecheck (push) Successful in 34s
CI & Build / integration (push) Successful in 34s
CI & Build / Build & push image (push) Has been skipped

Closes #2092 and the Knowledge-browse provenance gap.

The two halves of a hybrid search disagreed: the keyword half honoured shares
while the semantic half was pinned to NoteEmbedding.user_id, so a shared record
was findable by wording and invisible by meaning — the case a semantic search
exists to serve. semantic_search_notes now scopes on Note via a `scope`
parameter, and each of its five callers declares which kind of act it is:

  mcp/tools/search.py     read    the agent asked
  routes/search.py        read    the user typed it
  knowledge.py (semantic) read    matches the keyword half beside it
  plugin_context.py       browse  nobody asked; never a one-to-one share
  dedup.py                own     a verdict that blocks a write must not hinge
                                  on another person's notes

That last one is the reason this isn't a single global widening: the dedup gate
returns "update the existing one instead", so matching a stranger's record would
refuse a legitimate create and point at something the caller can't edit. Scope
defaults to "own" so a caller that forgets is wrong in the safe direction, and an
unknown scope raises rather than falling back — a typo there would be a
data-exposure bug.

Auto-inject keeps the browse scope, which still admits a collaborator's note via
a shared project. Its menu line is the only provenance an agent sees, so a
foreign hit now reads: #12 "Title" (0.71) - shared by alex, treat as a
suggestion. MCP and REST search results carry shared/owner too.

Knowledge browse: the feed hydrates cards from /api/knowledge/batch rather than
the list route, so both paths label rows now, and KnowledgeView shows "by
<owner>" on records the viewer doesn't own.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLwAaV4DQEmVyn496HnEvt
This commit is contained in:
2026-07-25 23:06:52 -04:00
parent 8b069cc93f
commit ef1dbdfc86
11 changed files with 319 additions and 22 deletions
+40 -7
View File
@@ -182,6 +182,45 @@ async def can_write_note(user_id: int, note_id: int) -> bool:
# you has not earned that standing, so it waits until you go looking for it.
# ---------------------------------------------------------------------------
def notes_visibility_clause(user_id: int, scope: str = "own"):
"""The one place a retrieval declares how far it may see.
Every path that returns notes picks a scope by what KIND of act it is:
"own" — the caller's records only. For machinery whose answer must not
depend on other people: the near-duplicate gate can't block a
create because a stranger wrote something similar, and can't
point at a record the caller cannot edit.
"browse" — own + project-reachable. For passive surfaces, where an
unrequested record would read as endorsed.
"read" — everything the ACL permits. For explicit acts: a typed search,
a fetch by id.
Defaults to the narrowest, so a new caller that forgets to choose is wrong in
the safe direction.
"""
if scope == "own":
return Note.user_id == user_id
if scope == "browse":
return browsable_notes_clause(user_id)
if scope == "read":
return readable_notes_clause(user_id)
raise ValueError(f"unknown note scope {scope!r} (own | browse | read)")
async def owner_names_for(user_ids: set[int]) -> dict[int, str]:
"""{user_id: username} in one query. Empty input costs nothing."""
if not user_ids:
return {}
async with async_session() as session:
rows = (
await session.execute(
select(User.id, User.username).where(User.id.in_(user_ids))
)
).all()
return {uid: name for uid, name in rows}
def _my_group_ids(user_id: int):
"""The caller's group ids as a SUBQUERY, not a fetched list.
@@ -267,13 +306,7 @@ async def label_shared_items(user_id: int, items: list[dict]) -> list[dict]:
}
if not foreign:
return items
async with async_session() as session:
rows = (
await session.execute(
select(User.id, User.username).where(User.id.in_(foreign))
)
).all()
names = {uid: name for uid, name in rows}
names = await owner_names_for(foreign)
for it in items:
owner_id = it.get("user_id")
if owner_id is not None and owner_id != user_id: