Additional MCP find examples improvements #31149
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit significantly enhances the `find_examples` MCP tool to improve its usability and the quality of its output, primarily from an LLM interaction perspective. Key changes include: 1. **Structured Output:** The tool's output schema is updated to return rich metadata for each example, including its `title`, `summary`, `keywords`, and `required_packages`. This allows the AI to present results more intelligently and check for prerequisites. 2. **Prescriptive Schema Descriptions:** The descriptions for both the input and output schemas have been rewritten to be more prescriptive. They now guide the AI on *how* and *why* to use specific fields, acting as a form of prompt engineering to elicit more precise queries and better-formatted responses. 3. **Code Refactoring:** The output schema definition has been moved to a module-level constant to improve code readability and consistency, reducing nesting in the main tool declaration.
This commit introduces a feature to mark and filter code examples that use experimental APIs, ensuring that the `find_examples` tool provides production-safe results by default. Key changes: - The example format now supports an optional `experimental: true` flag in the front matter. - The database schema is updated with an `experimental` column. Both the build-time and runtime database generators now parse and store this flag. - The `find_examples` tool's input schema is enhanced with an `includeExperimental` boolean flag, which defaults to `false`. - The query logic is updated to filter out experimental examples unless `includeExperimental` is explicitly set to `true`. - The schema description for the new flag includes a strong prescriptive guardrail, instructing the AI to warn the user when it shows an experimental example.
This commit enhances the `find_examples` tool by adding support for contextual snippets in the search results. Key changes: - The SQL query now uses the FTS5 `snippet()` function to generate a snippet of the matched content, with search terms highlighted. - The tool's output schema is updated with a new optional `snippet` field. - The description for the new `snippet` field is highly prescriptive, teaching the AI how to use snippets to enable more sophisticated and efficient agentic workflows, such as summarizing results for the user or performing internal evaluation before selecting the best result.
clydin
added
the
action: review
alan-agius4
approved these changes
Sep 9, 2025
alan-agius4
added
action: merge
|
This issue has been automatically locked due to inactivity. Read more about our automatic conversation locking policy. This action has been performed automatically by a bot. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
See individual commits.