Skip to content

armgit / xerj

Public Imported from GitHub
0

AI search engine and MCP server for bots and agents. One Rust binary indexes code, docs, logs, JSONL/CSV and PDFs for BM25 full-text, vector and hybrid search, code search without grep, agent memory and RAG. Elasticsearch-compatible API; Corpus Hub serves hundreds of reference memory corpora.

polish(md): SECURITY_MODEL: update body_targets/BodyShape line refs (#1542)

xerj-orga042f6d760c5
diff --git a/docs/SECURITY_MODEL.md b/docs/SECURITY_MODEL.md
index 7999f317d..521ed4630 100644
--- a/docs/SECURITY_MODEL.md
+++ b/docs/SECURITY_MODEL.md
@@ -48,15 +48,15 @@ It does four things:
`_mapping`, `_close`, `_alias`, `_forcemerge` and others, `authz.rs:441`)
need manage, as does a request with no operation segment at all, because that
creates or destroys the index. Everything else is treated as a write.
-3. Extracts the indices the request body names (`body_targets`, `authz.rs:730`)
+3. Extracts the indices the request body names (`body_targets`, `authz.rs:786`)
and authorizes each one. The shapes it parses are enumerated in `BodyShape`
- (`authz.rs:578`): `_bulk` action lines, `_msearch` header lines, `_mget`
+ (`authz.rs:634`): `_bulk` action lines, `_msearch` header lines, `_mget`
`docs[]._index`, `_aliases` actions, `_reindex` `source`/`dest`, snapshot
`indices`, `PUT /{index}` `aliases`, index-template `index_patterns`, ML job
and datafeed configs, the native `POST /v1/indices` body, and a `terms`
lookup or `lookup` runtime field at any depth inside a search body. Aliases
are resolved to their backing indices before the decision
- (`authorize_expression`, `authz.rs:1168`).
+ (`authorize_expression`, `authz.rs:1224`).
4. Installs the principal as the engine's visibility rule for the rest of the
request and prunes reserved index names out of metadata responses
(`prune_response`, `authz.rs:1431`).