Fix extension diagnostics and trust status reporting

AGENT_GUIDE.md

@@ -131,7 +131,7 @@ CI watching must be bounded. Do not loop indefinitely.
 
 ## Status Contract
 
-- `status --json` and related JSON/MCP payloads currently expose the trust fields documented in `README.md` and `DEVELOPER_GUIDE.md`, including `fold_ready`, `fold_ready_reason`, `graph_table_available`, `issues_table_available`, `file_issues_data_current`, `migration_in_progress`, `sql_graph_contract_ready`, `sql_graph_contract_degraded_reason`, `hotspot_family_ready`, `hotspot_family_degraded_reason`, `language_readiness`, `csharp_symbol_name_ready`, `csharp_metadata_target_ready`, `csharp_metadata_target_degraded_reason`, `indexed_head_commit`, `worktree_head_changed`, `index_writer_version`, `index_newer_than_reader`, `index_newer_than_reader_reason`, `unknown_extension_file_count`, `unknown_extension_files`, `unknown_extension_files_truncated`, `unknown_extension_file_path_limit`, `unknown_extension_extension_counts`, `unknown_extension_category_counts`, `unknown_extension_groups`, `extractors`, `path_case_sensitive`, `data_dir`, `data_dir_source`, `data_dir_mode`, `mac_profile`, `mac_profile_diagnostics`, `db_size_bytes`, `wal_size_bytes`, `db_pragma_settings` (`journal_mode`, `synchronous`, `wal_autocheckpoint`, `page_count`, `freelist_count`, `page_size`, `auto_vacuum`), `maintenance_guidance`, `symbol_kinds`, `symbols_by_language`, status kind cap metadata (`symbol_kind_limit`, `symbol_kind_name_limit`, `symbol_kind_total_count`, `symbol_kind_omitted_count`, `symbol_kind_names_truncated`, `symbols_by_language_kind_total_counts`, `symbols_by_language_kind_omitted_counts`, `symbols_by_language_kind_names_truncated`), `process`, `last_index_run`, `last_failed_or_partial_index_run`, `last_workspace_freshened_at`, `hooks`, `hook_diagnostics`, MCP-only `mcp_session`, and the `status --check`-only `stale_after_seconds` / `index_age_seconds` threshold audit fields and `repair_commands`.
+- `status --json` and related JSON/MCP payloads currently expose the trust fields documented in `README.md` and `DEVELOPER_GUIDE.md`, including `fold_ready`, `fold_ready_reason`, `graph_table_available`, `issues_table_available`, `file_issues_data_current`, `migration_in_progress`, `sql_graph_contract_ready`, `sql_graph_contract_degraded_reason`, `hotspot_family_ready`, `hotspot_family_degraded_reason`, `language_readiness`, `csharp_symbol_name_ready`, `csharp_metadata_target_ready`, `csharp_metadata_target_degraded_reason`, `indexed_head_commit`, `worktree_head_changed`, `index_writer_version`, `index_newer_than_reader`, `index_newer_than_reader_reason`, `unknown_extension_file_count`, `unknown_extension_files`, `unknown_extension_files_truncated`, `unknown_extension_file_path_limit`, `unknown_extension_extension_counts`, `unknown_extension_category_counts`, `unknown_extension_groups`, `extractors`, `path_case_sensitive`, `data_dir`, `data_dir_source`, `data_dir_mode`, `mac_profile`, `mac_profile_diagnostics`, `db_size_bytes`, `wal_size_bytes`, `db_pragma_settings` (`journal_mode`, `synchronous`, `wal_autocheckpoint`, `page_count`, `freelist_count`, `page_size`, `auto_vacuum`), `maintenance_guidance`, `symbol_kinds`, `symbols_by_language`, status kind cap metadata (`symbol_kind_limit`, `symbol_kind_name_limit`, `symbol_kind_total_count`, `symbol_kind_omitted_count`, `symbol_kind_names_truncated`, `symbols_by_language_kind_total_counts`, `symbols_by_language_kind_omitted_counts`, `symbols_by_language_kind_names_truncated`), `process`, `last_index_run`, `last_failed_or_partial_index_run`, `last_workspace_freshened_at`, `hooks`, `hook_diagnostics`, `trust_overrides`, MCP-only `mcp_session`, and the `status --check`-only `stale_after_seconds` / `index_age_seconds` threshold audit fields and `repair_commands`.
 - When any readiness field is degraded, the CLI adds `degraded_root_cause`, `degraded_reason`, `recommended_action`, `alternative_action`, and `readiness_degradations[]`. `degraded_root_cause` is the primary stable machine code; `readiness_degradations[]` lists every degraded field with `root_cause`, human reason, and remediation strings.
 - `hotspot_family_degraded_reason` currently uses `hotspot_family_support_not_indexed`, `hotspot_family_metadata_stale`, `hotspot_family_disabled_at_index_time`, `partial_family_key_population`, and `hotspot_family_marker_fingerprint_incomplete`; the incomplete marker fingerprint code means marker traversal hit safety caps and should stay synchronized with README / developer-guide recovery notes.
 - `issues_table_available` reports physical `file_issues` table presence only. `file_issues_data_current` reports whether the table is also stamped current for the active index generation.
@@ -140,6 +140,7 @@ CI watching must be bounded. Do not loop indefinitely.
 - `status` also surfaces unknown-extension scan coverage via `unknown_extension_file_count`, stamped by successful full-repository index runs (`cdidx index <projectPath>` and MCP `index_project`) as the number of non-indexed files with non-empty extensions that do not map to a known language. Current scans also stamp `unknown_extension_files` as a path sample bounded by `unknown_extension_file_path_limit` items and the string-list decoded-character budget, `unknown_extension_files_truncated` when more paths existed than were emitted for either bound, and `unknown_extension_file_path_limit` as the item cap rather than a guarantee that that many paths are returned. Newer scans also expose `unknown_extension_extension_counts`, `unknown_extension_category_counts`, and `unknown_extension_groups`; groups classify common non-code buckets such as repository metadata, licenses, binary assets, configuration, structural metadata, and language-support candidates, and include `recommended_action` values of `ignore_configuration`, `first_class_structural_extraction`, or `language_support`. These fields are omitted on legacy DBs or before a current full scan has stamped them.
 - `status` also surfaces extractor plugin and pattern-config runtime diagnostics via `extractors`, including loaded counts, skipped file counts, and a bounded diagnostics list for incompatible or malformed plugin/pattern files. Diagnostic paths, categories, and messages are sanitized before output; `diagnostics[].category` is the stable machine-readable failure code.
 - `status` also surfaces metadata-only post-extraction hook candidates and callback budgets through `hooks[]` / `hooks[].callback_budget_ms` without loading hook assemblies. `hook_diagnostics[]` reports sanitized hook discovery and callback diagnostics such as candidate-limit truncation, assembly load failure, constructor failure, callback failure, and timeout. `hook_diagnostics[].category` is the stable machine-readable failure code. Index runs still enforce `CDIDX_HOOK_CALLBACK_BUDGET_MS` (default: 5000 ms) on scratch copies, discard timed-out mutations, and disable timed-out hooks for the remainder of the current run.
+- `status` also surfaces accepted extension trust-boundary environment overrides through `trust_overrides[]`. Entries include `kind`, `environment_variable`, sanitized `value`, optional sanitized `path`, and `message`; current entries cover `CDIDX_TRUST_WORKSPACE_PLUGINS` workspace plugin discovery and `CDIDX_HOOKS_DIR` hook directory overrides.
 - `status` also surfaces `.cdidx` data-directory permissions via `data_dir_mode` on POSIX filesystems. New `.cdidx` data directories are forced to `0700`; the field is omitted on Windows, URI DBs, or when the directory mode cannot be inspected.
 - `status` also surfaces filesystem case-sensitivity via `path_case_sensitive`, stamped on every successful `cdidx index` run (full scan AND partial update, plus MCP-driven indexes) from `core.ignorecase` + a live filesystem probe. `true` means the volume is case-sensitive (`Foo.cs` and `foo.cs` are distinct); `false` means case-insensitive. Omitted on legacy DBs that predate the stamp. Use it to audit path-equality decisions on case-sensitive APFS, WSL NTFS / dev-drive, and ReFS mounts where the prior OS-keyed heuristic could mis-classify the workspace (#1546).
 - `status` also surfaces Linux mandatory-access-control context via `mac_profile` when `/proc/self/attr/current` or `/proc/self/attr/exec` indicates an AppArmor or SELinux profile. If proc attribute reads fail on Linux, `mac_profile_diagnostics[]` reports bounded `path`, `category`, and `message` entries so users can distinguish "no profile" from "profile detection failed" (#1768, #3480).

DEVELOPER_GUIDE.md

@@ -1058,7 +1058,7 @@ For the AI agent search-rule template, see [AI Integration](USER_GUIDE.md#ai-int
 | Readiness and graph trust | `fold_ready`, `fold_ready_reason`, `graph_table_available`, `issues_table_available`, `file_issues_data_current`, `migration_in_progress`, `sql_graph_contract_ready`, `sql_graph_contract_degraded_reason`, `hotspot_family_ready`, `hotspot_family_degraded_reason`, `language_readiness`, `csharp_symbol_name_ready`, `csharp_metadata_target_ready`, `csharp_metadata_target_degraded_reason`. |
 | Workspace and HEAD freshness | `indexed_head_commit`, `worktree_head_changed`, `indexed_head_sha`, `indexed_head_branch`, `indexed_head_timestamp`, `commits_ahead_of_indexed_head`. |
 | Version and forward compatibility | `index_writer_version`, `index_newer_than_reader`, `index_newer_than_reader_reason`. |
-| Unknown-extension and runtime diagnostics | `unknown_extension_file_count`, `unknown_extension_files`, `unknown_extension_files_truncated`, `unknown_extension_file_path_limit`, `unknown_extension_extension_counts`, `unknown_extension_category_counts`, `unknown_extension_groups`, `extractors`, `hooks`, `hook_diagnostics`, `path_case_sensitive`, `data_dir_mode`, `mac_profile`, `mac_profile_diagnostics`, `stale_after_seconds`, `index_age_seconds`, `last_failed_or_partial_index_run`. |
+| Unknown-extension and runtime diagnostics | `unknown_extension_file_count`, `unknown_extension_files`, `unknown_extension_files_truncated`, `unknown_extension_file_path_limit`, `unknown_extension_extension_counts`, `unknown_extension_category_counts`, `unknown_extension_groups`, `extractors`, `hooks`, `hook_diagnostics`, `trust_overrides`, `path_case_sensitive`, `data_dir_mode`, `mac_profile`, `mac_profile_diagnostics`, `stale_after_seconds`, `index_age_seconds`, `last_failed_or_partial_index_run`. |
 | Database maintenance | `db_size_bytes`, `wal_size_bytes`, `db_pragma_settings` (`journal_mode`, `synchronous`, `wal_autocheckpoint`, `page_count`, `freelist_count`, `page_size`, `auto_vacuum`), `maintenance_guidance`. |
 | Remediation fields | `degraded_root_cause`, `degraded_reason`, `recommended_action`, `alternative_action`, `readiness_degradations`, `repair_commands`. |
 | MCP-only session diagnostics | `mcp_session`, which is session-scoped diagnostics rather than persisted DB state. It contains `log_level`, bounded `roots`, optional `client_info`, and bounded optional `client_capabilities`. When advertised roots are capped, `roots_truncated`, `root_count`, `root_limit`, and `root_uri_length_limit` describe the truncation. When client capabilities are capped, `client_capabilities_truncated`, `client_capabilities_truncation_reason`, `client_capabilities_serialized_bytes`, `client_capabilities_byte_limit`, and `client_capabilities_depth_limit` describe the retained diagnostic subset. |
@@ -1073,6 +1073,7 @@ Runtime diagnostic subcontracts:
 | `hotspot_family_marker_fingerprint_incomplete` | Marker fingerprint traversal hit safety caps during the last index run; narrow or ignore generated/vendor marker trees before rebuilding. |
 | `extractors` | Reports runtime extractor plugin and pattern-config health, including loaded plugin assembly and pattern counts, symbol/reference extractor counts, skipped file counts, and a bounded diagnostic list for incompatible or malformed files. Diagnostic paths and messages are sanitized before output. |
 | `hooks[]` / `hook_diagnostics[]` | Includes metadata-only hook candidates and `callback_budget_ms`, mirroring the post-extraction callback budget enforced by `CDIDX_HOOK_CALLBACK_BUDGET_MS` (default: 5000 ms). `hook_diagnostics[]` reports sanitized discovery diagnostics such as candidate-limit truncation and `CDIDX_HOOKS_DIR` override acceptance or rejection without loading hook assemblies. Index runs still load hooks and discard timed-out callback mutations because hooks run on a scratch copy before their results are applied. |
+| `trust_overrides[]` | Reports accepted extension trust-boundary environment overrides, including `kind`, `environment_variable`, sanitized `value`, optional sanitized `path`, and `message`. Current entries cover accepted workspace plugin discovery via `CDIDX_TRUST_WORKSPACE_PLUGINS` and accepted hook directory overrides via `CDIDX_HOOKS_DIR`. |
 
 `references` already prefixes each human-readable row with `reference_kind`, and `callers` does the same for its grouped caller rows. When one grouped container mixes kinds (for example `call` and `subscribe` on the same event member), the human-readable label joins the distinct kinds with `+` (for example `call+subscribe`) instead of collapsing to a single preferred label, and the reference-kind column widens dynamically to fit the longest label in the batch so mixed rows do not overrun the neighbouring column. JSON output for `callers` and `callees` keeps the scalar `reference_kind` for back-compat (it reports the preferred summary kind `instantiate` > `subscribe` > `MIN(call)`) and adds a sorted `reference_kinds` array plus a `has_mixed_reference_kinds` bool so consumers can detect mixed containers without trusting a single collapsed label. This lets terminal users distinguish `call` / `instantiate` / `subscribe` / mixed without re-running the command with `--json` and lets AI clients answer mixed-kind questions without chasing a second `--exact` query.
 
@@ -3288,7 +3289,7 @@ AI エージェント向け検索ルールのテンプレートについては
 | readiness / graph trust | `fold_ready`, `fold_ready_reason`, `graph_table_available`, `issues_table_available`, `file_issues_data_current`, `migration_in_progress`, `sql_graph_contract_ready`, `sql_graph_contract_degraded_reason`, `hotspot_family_ready`, `hotspot_family_degraded_reason`, `language_readiness`, `csharp_symbol_name_ready`, `csharp_metadata_target_ready`, `csharp_metadata_target_degraded_reason`。 |
 | workspace / HEAD freshness | `indexed_head_commit`, `worktree_head_changed`, `indexed_head_sha`, `indexed_head_branch`, `indexed_head_timestamp`, `commits_ahead_of_indexed_head`。 |
 | version / forward compatibility | `index_writer_version`, `index_newer_than_reader`, `index_newer_than_reader_reason`。 |
-| unknown-extension / runtime diagnostics | `unknown_extension_file_count`, `unknown_extension_files`, `unknown_extension_files_truncated`, `unknown_extension_file_path_limit`, `unknown_extension_extension_counts`, `unknown_extension_category_counts`, `unknown_extension_groups`, `extractors`, `hooks`, `hook_diagnostics`, `path_case_sensitive`, `data_dir_mode`, `mac_profile`, `mac_profile_diagnostics`, `stale_after_seconds`, `index_age_seconds`, `last_failed_or_partial_index_run`。 |
+| unknown-extension / runtime diagnostics | `unknown_extension_file_count`, `unknown_extension_files`, `unknown_extension_files_truncated`, `unknown_extension_file_path_limit`, `unknown_extension_extension_counts`, `unknown_extension_category_counts`, `unknown_extension_groups`, `extractors`, `hooks`, `hook_diagnostics`, `trust_overrides`, `path_case_sensitive`, `data_dir_mode`, `mac_profile`, `mac_profile_diagnostics`, `stale_after_seconds`, `index_age_seconds`, `last_failed_or_partial_index_run`。 |
 | database maintenance | `db_size_bytes`, `wal_size_bytes`, `db_pragma_settings` (`journal_mode`, `synchronous`, `wal_autocheckpoint`, `page_count`, `freelist_count`, `page_size`, `auto_vacuum`), `maintenance_guidance`。 |
 | remediation fields | `degraded_root_cause`, `degraded_reason`, `recommended_action`, `alternative_action`, `readiness_degradations`, `repair_commands`。 |
 | MCP-only session diagnostics | `mcp_session`。これは persisted DB state ではなく session-scoped diagnostics で、`log_level`、上限付きの `roots`、任意の `client_info`、上限付きの任意の `client_capabilities` を含みます。advertised root が切り詰められた場合は `roots_truncated`、`root_count`、`root_limit`、`root_uri_length_limit` が切り詰め内容を示します。client capabilities が切り詰められた場合は `client_capabilities_truncated`、`client_capabilities_truncation_reason`、`client_capabilities_serialized_bytes`、`client_capabilities_byte_limit`、`client_capabilities_depth_limit` が保持された診断 subset を示します。 |
@@ -3303,6 +3304,7 @@ runtime diagnostic subcontract:
 | `hotspot_family_marker_fingerprint_incomplete` | 前回 index run で marker fingerprint traversal が safety cap に当たったことを示します。rebuild 前に generated / vendor marker tree を narrow または ignore してください。 |
 | `extractors` | runtime extractor plugin と pattern-config の health を報告します。loaded plugin assembly / pattern count、symbol/reference extractor count、skipped file count、incompatible / malformed file 用の bounded diagnostic list を含みます。diagnostic の path と message は出力前に sanitization されます。 |
 | `hooks[]` / `hook_diagnostics[]` | metadata-only の hook candidate と `callback_budget_ms` を含みます。これは `CDIDX_HOOK_CALLBACK_BUDGET_MS`（既定 5000 ms）が enforce する post-extraction callback budget と対応します。`hook_diagnostics[]` は candidate limit による打ち切りや `CDIDX_HOOKS_DIR` override の accepted / rejected など、sanitization 済みの discovery diagnostics を hook assembly を読み込まずに返します。index run は従来どおり hook を読み込んだうえで scratch copy 上の timeout した callback mutation を破棄します。 |
+| `trust_overrides[]` | 受理された拡張信頼境界の環境変数 override を報告します。各 entry は `kind`、`environment_variable`、sanitization 済みの `value`、任意の sanitization 済み `path`、`message` を含みます。現在は `CDIDX_TRUST_WORKSPACE_PLUGINS` による workspace plugin discovery と `CDIDX_HOOKS_DIR` による hook directory override の受理を対象にします。 |
 
 `references` は以前から人間向け出力の各行先頭に `reference_kind` を表示しており、`callers` も grouped caller 行に対して同じタグを出す。1 つの grouped container で kind が混在する場合（例: 同じ event メンバに対する `call` と `subscribe`）は、単一 preferred label へ潰さずに `call+subscribe` のように distinct kind を `+` で連結して表示する。reference-kind 列の幅はバッチ内で最も長いラベルに合わせて動的に広がるため、mixed 行が隣接列を押し出さない。`callers` / `callees` の JSON 出力では、後方互換のため scalar な `reference_kind`（preferred 順 `instantiate` > `subscribe` > `MIN(call)` の要約 kind）を残しつつ、ソート済みの `reference_kinds` 配列と `has_mixed_reference_kinds` bool も追加した。これにより consumer は単一 summary label に騙されずに mixed container を検出できる。端末上でも `call` / `instantiate` / `subscribe` / mixed を `--json` なしで見分けられ、AI クライアントも `--exact` を改めて投げ直さずに mixed-kind の問いに答えられる。