Fix/network filter coercion#494
Open
johndimaster-ai wants to merge 3 commits into
Open
Conversation
When an LLM calls search_people via MCP, it may pass the network parameter as a JSON-encoded string (e.g. '["F"]') rather than an actual Python list, causing a Pydantic validation error: 'Input should be a valid list [type=list_type]' Fix: add a coercion block at the top of search_people that detects a string network value and parses it as JSON, falling back to wrapping it in a list if parsing fails. Also widens the type hint to list[str] | str | None to reflect the accepted inputs. Handles all cases: '["F"]' -> ["F"] (JSON array string) '["F","S"]' -> ["F","S"] '"F"' -> ["F"] (JSON quoted string) 'F' -> ["F"] (bare string fallback) ["F"] -> ["F"] (already a list, no-op) None -> None (no filter, no-op)
fix: coerce network param from JSON string to list in search_people
The MCP framework validates tool parameters via Pydantic *before* the function body runs. The previous approach of coercing a JSON-serialised string (e.g. '["F"]') inside the function body therefore had no effect — Pydantic would already reject the string as an invalid list[str]. Fix: introduce a NetworkFilter type alias using Annotated + BeforeValidator that normalises any string input to list[str] before Pydantic's type check. The now-redundant manual coercion block in the function body is removed. Fixes: search_people called with network=['F'] failing with 'Input should be a valid list [type=list_type, input_value="["F"]", input_type=str]'
Contributor
Greptile SummaryThis PR moves
Confidence Score: 3/5This should be fixed before merging.
linkedin_mcp_server/tools/person.py Important Files Changed
Prompt To Fix All With AIFix the following 2 code review issues. Work through them one at a time, proposing concise fixes.
---
### Issue 1 of 2
linkedin_mcp_server/tools/person.py:138
**Coercion is bypassable**
`network` is now normalized only by the Pydantic validator. Callers that use the registered `FunctionTool.fn` directly, as the existing tests do, bypass that validator and pass the raw string into `extractor.search_people`. With `network='["F"]'`, the extractor iterates the string character by character and rejects tokens like `[` and `"`, so a previously accepted serialized network filter now fails instead of searching first-degree connections.
### Issue 2 of 2
linkedin_mcp_server/tools/person.py:39-40
**Serialized null fails**
The validator accepts JSON-serialized strings, but a serialized optional value of `"null"` parses to `None` and then gets wrapped as `[None]`. That makes an omitted network filter fail validation or extractor token checks instead of behaving the same as `network=None`.
```suggestion
parsed = _json.loads(v)
if parsed is None:
return None
return parsed if isinstance(parsed, list) else [parsed]
```
Reviews (1): Last reviewed commit: "fix(search_people): coerce network param..." | Re-trigger Greptile |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
| ctx: Context, | ||
| location: str | None = None, | ||
| network: list[str] | None = None, | ||
| network: NetworkFilter = None, |
Contributor
There was a problem hiding this comment.
network is now normalized only by the Pydantic validator. Callers that use the registered FunctionTool.fn directly, as the existing tests do, bypass that validator and pass the raw string into extractor.search_people. With network='["F"]', the extractor iterates the string character by character and rejects tokens like [ and ", so a previously accepted serialized network filter now fails instead of searching first-degree connections.
Prompt To Fix With AI
This is a comment left during a code review.
Path: linkedin_mcp_server/tools/person.py
Line: 138
Comment:
**Coercion is bypassable**
`network` is now normalized only by the Pydantic validator. Callers that use the registered `FunctionTool.fn` directly, as the existing tests do, bypass that validator and pass the raw string into `extractor.search_people`. With `network='["F"]'`, the extractor iterates the string character by character and rejects tokens like `[` and `"`, so a previously accepted serialized network filter now fails instead of searching first-degree connections.
How can I resolve this? If you propose a fix, please make it concise.
Comment on lines
+39
to
+40
| parsed = _json.loads(v) | ||
| return parsed if isinstance(parsed, list) else [parsed] |
Contributor
There was a problem hiding this comment.
The validator accepts JSON-serialized strings, but a serialized optional value of "null" parses to None and then gets wrapped as [None]. That makes an omitted network filter fail validation or extractor token checks instead of behaving the same as network=None.
Suggested change
| parsed = _json.loads(v) | |
| return parsed if isinstance(parsed, list) else [parsed] | |
| parsed = _json.loads(v) | |
| if parsed is None: | |
| return None | |
| return parsed if isinstance(parsed, list) else [parsed] |
Prompt To Fix With AI
This is a comment left during a code review.
Path: linkedin_mcp_server/tools/person.py
Line: 39-40
Comment:
**Serialized null fails**
The validator accepts JSON-serialized strings, but a serialized optional value of `"null"` parses to `None` and then gets wrapped as `[None]`. That makes an omitted network filter fail validation or extractor token checks instead of behaving the same as `network=None`.
```suggestion
parsed = _json.loads(v)
if parsed is None:
return None
return parsed if isinstance(parsed, list) else [parsed]
```
How can I resolve this? If you propose a fix, please make it concise.
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
No description provided.