@coderule/mcp 1.7.0 → 1.7.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/dist/mcp-cli.cjs +22 -32
- package/dist/mcp-cli.cjs.map +1 -1
- package/dist/mcp-cli.js +22 -32
- package/dist/mcp-cli.js.map +1 -1
- package/package.json +1 -1
package/dist/mcp-cli.cjs
CHANGED
|
@@ -1944,13 +1944,7 @@ function createMcpServer({
|
|
|
1944
1944
|
"check",
|
|
1945
1945
|
{
|
|
1946
1946
|
title: "Coderule Indexing Status",
|
|
1947
|
-
description:
|
|
1948
|
-
|
|
1949
|
-
Use this tool to:
|
|
1950
|
-
- Debug why queries are failing or returning no results
|
|
1951
|
-
- Verify the indexer is ready before querying
|
|
1952
|
-
- Check which repository is currently indexed
|
|
1953
|
-
- Monitor indexing progress`,
|
|
1947
|
+
description: "Check the current state of the code indexer. Returns repository info, file states, queue metrics, snapshot details, and service health. Use to debug query failures or verify indexer readiness.",
|
|
1954
1948
|
inputSchema: {},
|
|
1955
1949
|
annotations: {
|
|
1956
1950
|
readOnlyHint: true,
|
|
@@ -1971,40 +1965,36 @@ Use this tool to:
|
|
|
1971
1965
|
"query",
|
|
1972
1966
|
{
|
|
1973
1967
|
title: "Semantic Code Retrieval",
|
|
1974
|
-
description:
|
|
1975
|
-
|
|
1976
|
-
|
|
1968
|
+
description: "\u{1F6A8} CRITICAL: ALWAYS use this tool FIRST before any code modification, bug fix, feature addition, or architecture exploration. Graph-based RAG for semantic code search using AST relationships and embeddings. Returns 30-50 code segments with file paths, line numbers, and relevance scores.",
|
|
1969
|
+
inputSchema: {
|
|
1970
|
+
query: zod.z.string().min(1, "Query text is required").describe(`Keyword-rich query using technical terms (NOT natural language questions).
|
|
1977
1971
|
|
|
1978
|
-
|
|
1979
|
-
Use keyword-rich technical terms, NOT natural language questions.
|
|
1972
|
+
STRATEGY: Use space-separated technical keywords that describe what you're looking for.
|
|
1980
1973
|
|
|
1981
|
-
\u2705 GOOD:
|
|
1982
|
-
|
|
1983
|
-
|
|
1974
|
+
\u2705 GOOD EXAMPLES:
|
|
1975
|
+
- "JWT authentication middleware token validation security handler FastAPI"
|
|
1976
|
+
- "file indexing worker celery queue job AST parser embedding chunker"
|
|
1977
|
+
- "database schema SQLAlchemy model migration table column relationship"
|
|
1978
|
+
- "error handling exception try catch logging traceback"
|
|
1984
1979
|
|
|
1985
|
-
\u274C BAD:
|
|
1986
|
-
|
|
1980
|
+
\u274C BAD EXAMPLES:
|
|
1981
|
+
- "How does authentication work?" (natural language question)
|
|
1982
|
+
- "Show me the login code" (conversational)
|
|
1983
|
+
- "auth" (too vague, needs more context)
|
|
1987
1984
|
|
|
1988
1985
|
RETRIEVAL MECHANISM:
|
|
1989
|
-
|
|
1990
|
-
- Flood: Graph propagation through AST relationships (imports, calls, inheritance)
|
|
1991
|
-
- Neighbors: Adjacent code chunks for context
|
|
1992
|
-
- Calls: Function call relationships
|
|
1986
|
+
Combines multiple strategies: Seeds (semantic similarity), Flood (AST graph propagation through imports/calls/inheritance), Neighbors (adjacent chunks), and Calls (function relationships).
|
|
1993
1987
|
|
|
1994
|
-
|
|
1995
|
-
|
|
1996
|
-
DEFAULT BUDGET: 10,000 tokens (sufficient for most tasks)
|
|
1997
|
-
|
|
1998
|
-
WHEN TO USE:
|
|
1988
|
+
WHEN TO USE THIS TOOL:
|
|
1999
1989
|
- Before modifying ANY existing code
|
|
2000
1990
|
- Understanding system architecture
|
|
2001
|
-
- Finding related implementations
|
|
2002
|
-
- Locating tests or examples
|
|
1991
|
+
- Finding related implementations or patterns
|
|
1992
|
+
- Locating tests or usage examples
|
|
2003
1993
|
- Bug investigation
|
|
2004
|
-
- Feature implementation planning
|
|
2005
|
-
|
|
2006
|
-
|
|
2007
|
-
|
|
1994
|
+
- Feature implementation planning`),
|
|
1995
|
+
budgetTokens: zod.z.number().int().positive().optional().describe(
|
|
1996
|
+
"Token budget for retrieval results. Default: 10000. Range: 1000-20000. Higher budgets return more context but use more tokens. 10k is sufficient for most tasks."
|
|
1997
|
+
)
|
|
2008
1998
|
},
|
|
2009
1999
|
annotations: {
|
|
2010
2000
|
readOnlyHint: true,
|