Skip to content

Codebase intelligence for AI and humans: code health scores, auto-generated docs, git analytics, dead code detection, and architectural decisions via MCP.

fix(mcp): name the capability, not a host's tools, in no-index guidance (#2326)

Nehal kanak kuniyil thottathil32d00025423f
diff --git a/packages/server/src/repowise/server/mcp_server/_failure_shield.py b/packages/server/src/repowise/server/mcp_server/_failure_shield.py
index 73e00948d..fa85fa789 100644
--- a/packages/server/src/repowise/server/mcp_server/_failure_shield.py
+++ b/packages/server/src/repowise/server/mcp_server/_failure_shield.py
@@ -56,8 +56,9 @@ def _shape_not_indexed() -> dict[str, Any]:
),
"guidance": (
"Until an index exists, every repowise tool will return this "
- "notice. Answer questions about this repo with your built-in "
- "tools (Read/Grep/Glob) for the rest of the session."
+ "notice. Answer questions about this repo with your own "
+ "file-reading and searching tools, whatever they are called, "
+ "for the rest of the session."
),
}
@@ -132,8 +133,8 @@ def _shape_stale_index(exc: Exception) -> dict[str, Any]:
"guidance": (
"If the retry returns this notice too, every repowise tool will "
"keep returning it until the index is rebuilt. Answer questions "
- "about this repo with your built-in tools (Read/Grep/Glob) for the "
- "rest of the session."
+ "about this repo with your own file-reading and searching tools, "
+ "whatever they are called, for the rest of the session."
),
}
diff --git a/plugins/claude-code/skills/codebase-exploration/SKILL.md b/plugins/claude-code/skills/codebase-exploration/SKILL.md
index 433d7179b..ed0d14748 100644
--- a/plugins/claude-code/skills/codebase-exploration/SKILL.md
+++ b/plugins/claude-code/skills/codebase-exploration/SKILL.md
@@ -38,8 +38,8 @@ plain file reads don't, usually in one round-trip instead of many.
4. More files around a concept → `search_codebase`, then `get_context` on the
hits (batched), then `get_symbol` only for the bodies you actually need.
-Fall back to raw Read/Grep only when the indexed context doesn't cover the
-specific detail the user asked about.
+Fall back to your own file-reading and searching tools only when the indexed
+context doesn't cover the specific detail the user asked about.
## Trust signals — verify when
diff --git a/plugins/codex/skills/codebase-exploration/SKILL.md b/plugins/codex/skills/codebase-exploration/SKILL.md
index 0727f96fb..38020b78c 100644
--- a/plugins/codex/skills/codebase-exploration/SKILL.md
+++ b/plugins/codex/skills/codebase-exploration/SKILL.md
@@ -33,8 +33,8 @@ plain file reads don't, usually in one round-trip instead of many.
4. More files around a concept → `search_codebase`, then `get_context` on the
hits (batched), then `get_symbol` only for the bodies you actually need.
-Fall back to raw Read/Grep only when the indexed context doesn't cover the
-specific detail the user asked about.
+Fall back to your own file-reading and searching tools only when the indexed
+context doesn't cover the specific detail the user asked about.
## Trust signals — verify when
diff --git a/plugins/shared/skills/codebase-exploration.md b/plugins/shared/skills/codebase-exploration.md
index f0b0ebb36..c93cda767 100644
--- a/plugins/shared/skills/codebase-exploration.md
+++ b/plugins/shared/skills/codebase-exploration.md
@@ -46,8 +46,8 @@ plain file reads don't, usually in one round-trip instead of many.
4. More files around a concept → `search_codebase`, then `get_context` on the
hits (batched), then `get_symbol` only for the bodies you actually need.
-Fall back to raw Read/Grep only when the indexed context doesn't cover the
-specific detail the user asked about.
+Fall back to your own file-reading and searching tools only when the indexed
+context doesn't cover the specific detail the user asked about.
## Trust signals — verify when
diff --git a/tests/unit/server/mcp/test_failure_shield.py b/tests/unit/server/mcp/test_failure_shield.py
index f4cbd7824..85215da0b 100644
--- a/tests/unit/server/mcp/test_failure_shield.py
+++ b/tests/unit/server/mcp/test_failure_shield.py
@@ -43,8 +43,10 @@ async def test_uninitialized_repo_is_success_shaped(empty_mcp):
assert "repowise init" in result["remedy"]
# The user decides whether to index; the agent must not run init itself.
assert "user" in result["remedy"]
- # Session-scoped guidance: use built-in tools instead of retrying forever.
- assert "Read/Grep/Glob" in result["guidance"]
+ # Session-scoped guidance: fall back to the agent's own tools, named by
+ # capability rather than by one host's tool names.
+ assert "file-reading and searching tools" in result["guidance"]
+ assert "Read/Grep/Glob" not in result["guidance"]
@pytest.mark.asyncio
@@ -99,6 +101,10 @@ async def test_a_store_older_than_the_models_says_run_update_not_give_up():
assert "repowise update" in result["remedy"]
assert "user" in result["remedy"]
assert "Retry this call once" not in result.get("guidance", "")
+ # The same capability wording as the not-indexed shape; this is the second
+ # of the two sites, and a host tool name here would not travel either.
+ assert "file-reading and searching tools" in result["guidance"]
+ assert "Read/Grep/Glob" not in result["guidance"]
@pytest.mark.asyncio