New optimisations#472
Conversation
feat(mcp): add ChatGPT and outreach support
Greptile SummaryThis PR adds remote MCP support and new read-only workflow tools. The main changes are:
Confidence Score: 5/5This looks safe to merge.
Reviews (2): Last reviewed commit: "chore: ignore mcp bundle artifacts" | Re-trigger Greptile |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
| url = _absolute_linkedin_url(value) | ||
| extracted = await extractor.extract_page(url, section_name="post") |
There was a problem hiding this comment.
fetch accepts post:<value>, turns absolute URLs into themselves, and then navigates the authenticated browser to that URL. A client can call fetch with post:https://attacker.example/, which makes this LinkedIn fetch tool browse arbitrary external pages instead of only LinkedIn post permalinks. Please validate that absolute post URLs use a LinkedIn host and the expected /feed/update/ path before passing them to extract_page.
Prompt To Fix With AI
This is a comment left during a code review.
Path: linkedin_mcp_server/tools/compat.py
Line: 214-215
Comment:
**Restrict fetched post URLs**
`fetch` accepts `post:<value>`, turns absolute URLs into themselves, and then navigates the authenticated browser to that URL. A client can call `fetch` with `post:https://attacker.example/`, which makes this LinkedIn fetch tool browse arbitrary external pages instead of only LinkedIn post permalinks. Please validate that absolute post URLs use a LinkedIn host and the expected `/feed/update/` path before passing them to `extract_page`.
How can I resolve this? If you propose a fix, please make it concise.| return { | ||
| "schedule": [f"Day {day}" for day in days], | ||
| "drafts": drafts[: len(days)], |
There was a problem hiding this comment.
schedule includes every requested cadence day, but drafts is sliced from a fixed three-message list. When a client passes cadence_days=[1, 3, 7, 14], the response contains four schedule entries and only three drafts, so callers cannot pair each follow-up day with a message. Either cap or validate the cadence length, or generate one draft per requested day.
Prompt To Fix With AI
This is a comment left during a code review.
Path: linkedin_mcp_server/tools/outreach.py
Line: 235-237
Comment:
**Keep plans aligned**
`schedule` includes every requested cadence day, but `drafts` is sliced from a fixed three-message list. When a client passes `cadence_days=[1, 3, 7, 14]`, the response contains four schedule entries and only three drafts, so callers cannot pair each follow-up day with a message. Either cap or validate the cadence length, or generate one draft per requested day.
How can I resolve this? If you propose a fix, please make it concise.| def _trim_to_chars(text: str, max_chars: int) -> str: | ||
| compact = " ".join(text.split()) | ||
| if len(compact) <= max_chars: | ||
| return compact | ||
| return compact[: max_chars - 1].rstrip() + "..." |
There was a problem hiding this comment.
max_chars is accepted without a lower bound, but this helper assumes it is positive. If a caller passes max_chars=0 or a negative value, the slice uses max_chars - 1 and returns a long string with ..., while the response still reports the requested limit. Tiny positive values can also return a string longer than the limit. Add input validation for max_chars or handle small bounds explicitly so draft and connection-note length guarantees hold.
Prompt To Fix With AI
This is a comment left during a code review.
Path: linkedin_mcp_server/tools/outreach.py
Line: 75-79
Comment:
**Validate character limits**
`max_chars` is accepted without a lower bound, but this helper assumes it is positive. If a caller passes `max_chars=0` or a negative value, the slice uses `max_chars - 1` and returns a long string with `...`, while the response still reports the requested limit. Tiny positive values can also return a string longer than the limit. Add input validation for `max_chars` or handle small bounds explicitly so draft and connection-note length guarantees hold.
How can I resolve this? If you propose a fix, please make it concise.1 participant
No description provided.