weavatrix 1.17.4 → 1.17.6

package/skill/references/tools/get_community.md

@@ -11,7 +11,7 @@ Inspect one deterministic weak graph component by its id.
 - `community_id` (integer, required, min 0).
 - `cursor` (string).
 - `max_nodes` (integer, min 0).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 
@@ -25,5 +25,4 @@ Inspect one deterministic weak graph component by its id.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/get_dependents.md

@@ -13,7 +13,7 @@ Measure a bounded transitive reverse blast radius from one exact node.
 - `include_container_importers` (boolean).
 - `max_nodes` (integer, min 0).
 - `max_references` (integer, min 0).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `precision` ("graph").
 - `timeout_ms` (integer, min 0).
 
@@ -29,5 +29,4 @@ Measure a bounded transitive reverse blast radius from one exact node.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/get_neighbors.md

@@ -13,7 +13,7 @@ For transitive reverse impact use `get_dependents`; for bounded multi-hop explor
 - `label` (string, required).
 - `cursor` (string).
 - `max_results` (integer, min 0).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `relation_filter` (string | array<string>).
 - `response_detail` ("compact" | "full").
 
@@ -29,5 +29,4 @@ For transitive reverse impact use `get_dependents`; for bounded multi-hop explor
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/get_node.md

@@ -9,7 +9,7 @@ Resolve one exact graph label after another tool has supplied an unambiguous fil
 ## Inputs
 
 - `label` (string, required).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 
@@ -23,5 +23,4 @@ Resolve one exact graph label after another tool has supplied an unambiguous fil
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/git_history.md

@@ -4,7 +4,7 @@ Bounded direct Git history without launching git.
 
 ## When to use
 
-Inspect bounded commit history, churn, and co-change inside the active repository.
+Inspect bounded commit history inside the active repository. Without `path` this is repository HEAD history. With `path` it is the commits that touch that file or directory. The default six-month window applies only when `path` is omitted or `months` is set.
 
 ## Inputs
 
@@ -14,7 +14,8 @@ Inspect bounded commit history, churn, and co-change inside the active repositor
 - `max_pairs` (integer, min 0).
 - `min_pair_count` (integer, min 0).
 - `months` (integer, min 0).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `path` (string) — repository-relative file or directory. Omit it for repository HEAD history.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `revision` (string).
 - `token_budget` (integer, min 1) — Approximate output ceiling in tokens (serialized bytes / 4); result arrays are trimmed from the tail to fit and the report states what was dropped.
 - `top_n` (integer, min 0).
@@ -30,5 +31,4 @@ Inspect bounded commit history, churn, and co-change inside the active repositor
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/git_read_blob.md

@@ -10,7 +10,7 @@ After `graph_diff` or `git_history`, read a file "as it was" at a revision - or
 
 - `max_bytes` (integer, min 1) — cap on returned content bytes (default 262144); the true size and a truncation flag are always reported.
 - `oid` (string) — blob object id; mutually exclusive with `path`.
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `path` (string) — repository-relative file path looked up in the revision's tree.
 - `revision` (string) — any resolvable revision, `~N` first-parent steps supported; default `HEAD`.
 - `token_budget` (integer, min 1) — Approximate output ceiling in tokens (serialized bytes / 4); result arrays are trimmed from the tail to fit and the report states what was dropped.

package/skill/references/tools/go_to_definition.md

@@ -13,7 +13,7 @@ graph recorded for that occurrence - not a same-named overload elsewhere.
 - `line` (integer, required, min 1) — 1-based line of the usage.
 - `column` (integer, required, min 1) — 1-based column of the usage.
 - `scip_path` (string) — optional on-disk SCIP index; never spawned.
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 

package/skill/references/tools/god_nodes.md

@@ -9,7 +9,7 @@ Find unusually connected production nodes for architecture or change-risk review
 ## Inputs
 
 - `include_classified` (boolean).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `top_n` (integer, min 0).
 
 ## Minimal call
@@ -23,5 +23,4 @@ Find unusually connected production nodes for architecture or change-risk review
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/graph_diff.md

@@ -12,7 +12,7 @@ Compare the current graph snapshot with an immutable Git revision.
 - `detail` ("file_pairs" | "edges", default "file_pairs") — Aggregate edge churn by source file, target file, and relation by default; request edges for individual edge provenance.
 - `head_ref` (string).
 - `max_results` (integer, min 0).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `path` (string).
 - `token_budget` (integer, min 1) — Approximate output ceiling in tokens (serialized bytes / 4); result arrays are trimmed from the tail to fit and the report states what was dropped.
 
@@ -28,5 +28,4 @@ Compare the current graph snapshot with an immutable Git revision.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/graph_stats.md

@@ -9,7 +9,7 @@ Confirm the active repository, graph revision, freshness, counts, and available
 ## Inputs
 
 - `include_capabilities` (boolean).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 
@@ -22,5 +22,4 @@ Confirm the active repository, graph revision, freshness, counts, and available
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/hot_path_review.md

@@ -14,7 +14,7 @@ Rank static performance-review candidates using connectivity and change signals.
 - `include_tests` (boolean).
 - `loop_depth_threshold` (integer, min 0).
 - `min_score` (integer, min 0).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `path` (string).
 - `time_rank_threshold` (string).
 - `top_n` (integer, min 0).
@@ -30,5 +30,4 @@ Rank static performance-review candidates using connectivity and change signals.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/inspect_symbol.md

@@ -12,7 +12,7 @@ Collect one symbol declaration, owner, callers, callees, and source evidence.
 - `context_lines` (integer, min 0).
 - `max_containers` (integer, min 0).
 - `max_references` (integer, min 0).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `precision` (string).
 - `timeout_ms` (integer, min 0).
 
@@ -28,5 +28,4 @@ Collect one symbol declaration, owner, callers, callees, and source evidence.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/list_communities.md

@@ -9,7 +9,7 @@ List the largest deterministic weak graph components before choosing one to insp
 ## Inputs
 
 - `include_non_product` (boolean).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `top_n` (integer, min 0).
 
 ## Minimal call
@@ -23,5 +23,4 @@ List the largest deterministic weak graph components before choosing one to insp
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/list_endpoints.md

@@ -11,7 +11,7 @@ Inventory statically extracted HTTP endpoints with optional method and path filt
 - `include_classified` (boolean).
 - `max_results` (integer, min 0).
 - `method` (string).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `path` (string).
 
 ## Minimal call
@@ -25,5 +25,4 @@ Inventory statically extracted HTTP endpoints with optional method and path filt
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/list_known_repos.md

@@ -8,7 +8,7 @@ See which roots this process has already opened.
 
 ## Inputs
 
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 
@@ -21,5 +21,4 @@ See which roots this process has already opened.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/map_stacktrace.md

@@ -10,7 +10,7 @@ Map pasted Node, JVM, Python, or Rust stack frames to repository files and neare
 
 - `text` (string, required).
 - `max_frames` (integer, min 0).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 
@@ -24,5 +24,4 @@ Map pasted Node, JVM, Python, or Rust stack frames to repository files and neare
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/module_map.md

@@ -9,7 +9,7 @@ Orient in an unfamiliar repository through production folders and their dependen
 ## Inputs
 
 - `include_non_product` (boolean).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `top_n` (integer, min 0).
 
 ## Minimal call
@@ -23,5 +23,4 @@ Orient in an unfamiliar repository through production folders and their dependen
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/n8n_context.md

@@ -14,7 +14,7 @@ not placed on the default context.
 - `label` (string, required) — exact node or workflow identity.
 - `task` (string, optional) — what the caller intends to change or inspect.
 - `max_related` (number, optional) — related-item bound, default 24.
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 
@@ -29,5 +29,4 @@ not placed on the default context.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/n8n_inventory.md

@@ -11,7 +11,7 @@ does not call a live n8n API.
 
 - `path` (string, optional) — repository-relative workflow file or path fragment.
 - `max_results` (number, optional) — inventory page size, 1–500, default 200.
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 
@@ -24,5 +24,4 @@ does not call a live n8n API.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/n8n_trace.md

@@ -15,7 +15,7 @@ walk. Missing subworkflows are not provided, not deleted.
 - `depth` (number, optional) — hop bound, 1–32, default 8.
 - `max_nodes` (number, optional) — page size, 1–500, default 100.
 - `cursor` (string, optional) — opaque `v1:<offset>` token from `page.next_cursor`.
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 
@@ -29,5 +29,4 @@ walk. Missing subworkflows are not provided, not deleted.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/open_repo.md

@@ -1,17 +1,17 @@
 # `open_repo`
 
-Retarget to another local repository.
+Retarget only when this process was started with `--allow-retarget`.
 
 ## When to use
 
-Retarget the process-local server to another local repository.
+The launch root is fixed for the process lifetime unless `--allow-retarget` was passed. A pinned process refuses `open_repo` to another directory. Relaunch on that repository instead. A mismatched `expected_repository` is the same situation: relaunch, do not call `open_repo` on a pinned process.
 
 ## Inputs
 
 - `path` (string, required).
 - `build` (boolean).
 - `mode` ("full" | "no-tests" | "tests-only").
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `precision` (string).
 
 ## Minimal call
@@ -26,5 +26,4 @@ Retarget the process-local server to another local repository.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/perf_attribution.md

@@ -19,7 +19,7 @@ co-occurrence with structural change - not a profiler sample.
 - `path` (string).
 - `include_tests` (boolean).
 - `include_classified` (boolean).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 

package/skill/references/tools/prepare_change.md

@@ -10,7 +10,7 @@ Select the architecture rules relevant to intended changed files before editing.
 
 - `files` (array<string>, required).
 - `intent` (string).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 
@@ -24,5 +24,4 @@ Select the architecture rules relevant to intended changed files before editing.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/propose_architecture_exception.md

@@ -11,7 +11,7 @@ Prepare a reviewable exception proposal without writing it to the repository.
 - `fingerprint` (string, required).
 - `reason` (string, required).
 - `expires` (string).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 
@@ -26,5 +26,4 @@ Prepare a reviewable exception proposal without writing it to the repository.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/query_graph.md

@@ -18,7 +18,7 @@ Start with exact seeds and a small depth, node cap, and `token_budget`.
 - `include_low_signal` (boolean).
 - `max_nodes` (integer, min 0).
 - `mode` ("bfs" | "dfs").
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `question` (string).
 - `relation_filter` (string | array<string>).
 - `seed_files` (array<string>).
@@ -36,5 +36,4 @@ Start with exact seeds and a small depth, node cap, and `token_budget`.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/read_source.md

@@ -11,7 +11,7 @@ Read an exact verified line range after a file and location are known.
 - `after` (integer, min 0).
 - `before` (integer, min 0).
 - `label` (string).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `path` (string).
 - `start_line` (integer, min 0).
 - `token_budget` (integer, min 1) — Approximate output ceiling in tokens (serialized bytes / 4); result arrays are trimmed from the tail to fit and the report states what was dropped.
@@ -27,5 +27,4 @@ Read an exact verified line range after a file and location are known.
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/rebuild_graph.md

@@ -11,7 +11,7 @@ Do not rebuild on every call; the server normally refreshes derived state automa
 ## Inputs
 
 - `mode` ("full" | "no-tests" | "tests-only").
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `precision` (string).
 - `scope` (string).
 
@@ -26,5 +26,4 @@ Do not rebuild on every call; the server normally refreshes derived state automa
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/run_audit.md

@@ -16,7 +16,7 @@ Run a broad repository health and evidence-completeness pass before targeted inv
 - `include_classified` (boolean).
 - `max_findings` (integer, min 0).
 - `min_severity` ("low" | "medium" | "high" | "critical").
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 
 ## Minimal call
 
@@ -29,5 +29,4 @@ Run a broad repository health and evidence-completeness pass before targeted inv
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/search_code.md

@@ -4,7 +4,7 @@ Literal or Rust-regex repository search without ripgrep.
 
 ## When to use
 
-Find bounded literal or regular-expression matches when the source text is known but its location is not.
+Find bounded literal or regular-expression matches in file text when the source text is known but its location is not. The process keeps a resident content index: the first search builds it and later searches read it (`backend` `liveindex`). A `glob` still walks the filesystem. This is not a symbol-graph search. Use `find_references` or `go_to_definition` for a symbol.
 
 ## Inputs
 
@@ -14,7 +14,7 @@ Find bounded literal or regular-expression matches when the source text is known
 - `glob` (string).
 - `is_regex` (boolean).
 - `max_results` (integer, min 0).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `token_budget` (integer, min 1) — Approximate output ceiling in tokens (serialized bytes / 4); result arrays are trimmed from the tail to fit and the report states what was dropped.
 
 ## Minimal call
@@ -29,5 +29,4 @@ Find bounded literal or regular-expression matches when the source text is known
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.

package/skill/references/tools/select_tests.md

@@ -16,7 +16,7 @@ Rank the test suites most plausibly affected by a change; it selects tests but d
 - `head_ref` (string).
 - `max_nodes` (integer, min 0).
 - `max_tests` (integer, min 0).
-- `output_format` ("text" | "json" | "structured", default "json") — text returns the concise text block only; json returns structured output and mirrors it into text for clients that read only content; structured drops that mirror, which is the larger copy, and is safe only where the client reads structuredContent.
+- `output_format` ("text" | "json" | "structured", default "json") — text returns one JSON text block of the full payload, not a shorter summary; json returns structured output and mirrors that JSON into text, so the answer is about twice as large; structured drops the mirror and is the smaller payload when the client reads structuredContent.
 - `precision` ("graph").
 
 ## Minimal call
@@ -30,5 +30,4 @@ Rank the test suites most plausibly affected by a change; it selects tests but d
 }
 ```
 
-Use `output_format:"text"` for compact agent interaction. The live MCP
-`tools/list` schema remains authoritative for this installed version.
+Prefer `output_format:"structured"` when the client reads structuredContent. `text` is still the full JSON payload. The live MCP `tools/list` schema remains authoritative for this installed version.