# Turn LinkedIn bug reports into issues with Devin Source: https://finchreach.com/blog/devin-linkedin-bug-reports Published 2026-10-28 by Finchreach team. Tags: devin, linkedin, mcp, developer-tools, use-case. Have Devin find bug reports and feature requests about your dev tool on LinkedIn, file them as issues, and queue a thank-you for your approval. ## The short answer If you build a developer tool, people report bugs and ask for features in public on LinkedIn: under the founder’s launch posts, in their own posts about your product, in passing mentions. Connect Devin to Finchreach and it can check the founder’s LinkedIn notifications, search posts that mention your product, pick out the bug reports and feature requests, and turn them into issues in your repo or tracker, which is already Devin’s normal job. Devin can then draft a short thank-you as a comment on the post, or a direct message to the person who reported it. Neither goes out on its own. Approval mode is on by default for every Finchreach account, so each draft waits in the approval queue until a person approves or rejects it. The link between them is MCP (Model Context Protocol), a standard plug that lets AI apps use outside tools. Finchreach is a remote MCP server at `https://mcp.finchreach.com/mcp` that operates one LinkedIn account, authenticated with an API key. Devin supports remote MCP servers with an auth header, so it connects directly. ## Why collect bug reports from LinkedIn at all? Because that is where some users talk, and they rarely cross-post to your issue tracker. A comment like “tried it on our monorepo, the CLI hangs on install” is a real bug report, but it lives under a post, gets a few likes and then scrolls away. Nobody on the team files it, and the person who wrote it never hears back. Picking these up by hand is tedious, so it tends not to happen. It is a good fit for an agent: read, classify, check for duplicates, write a clear issue with a link to the source, and draft a reply. The parts that need judgment, whether the issue is real and what to say in public, stay with people. ## What can Devin see on LinkedIn through Finchreach? - `linkedin_get_notifications` returns the account’s latest notifications and unseen counts, which is where new comments and mentions on the founder’s posts show up. - `linkedin_search` with the `content` vertical finds posts mentioning your product name, with cursor paging to read further. - `linkedin_get_profile` reads a profile, so Devin can tell whether a reporter works at a customer, a prospect or nowhere related. - `linkedin_get_conversations` and `linkedin_get_messages` read the inbox, for reports that arrive as direct messages. Be aware of a gap: Finchreach has no tool that lists every comment on a given post. Devin sees comments through notifications and through search, which works well for recent activity but can miss older or quieter threads. If you know of a specific comment, paste its text and the post link into the session. ## What do you ask Devin? A first session can cover the whole loop: > Ask your assistant: “Check the LinkedIn notifications on the Finchreach-connected account and search LinkedIn posts from the last two weeks that mention "Acme CLI". List every comment or post that describes a bug or asks for a feature. For each, give the author, a quote, the post link, and whether it is a bug or a feature request.” Then hand over the part Devin already does well: > Ask your assistant: “For each bug, search our GitHub issues for duplicates. If there is none, open a new issue with a clear title, steps to reproduce as far as we can infer them, the quote, and the LinkedIn link as the source. Label feature requests "feature-request" instead of "bug". Show me the list of issues you created.” Finally, close the loop with the people who reported them: > Ask your assistant: “For the three clearest bug reports, draft a short thank-you comment on the original post: thank them, say we have opened an issue, and give the public issue link if the repo is public. Keep each under 300 characters, no marketing. Queue them through Finchreach and give me the action ids.” Each comment comes back with status `pending_approval`. Devin can poll it with `linkedin_get_action`, and the founder or whoever owns the account approves or rejects it at [/approvals](https://finchreach.com/approvals). > Comments from `linkedin_comment_on_post` are top-level comments on the post, not threaded replies under a specific comment. Mention the person by name in the text so the context is clear, or send a direct message with `linkedin_send_dm` when a private follow-up fits better. ## How do you connect Devin to LinkedIn? 1. Create a Finchreach account at [finchreach.com/sign-up](https://finchreach.com/sign-up), connect the founder’s LinkedIn account under Accounts, and create an API key under API keys. 2. In Devin, open Customize → MCPs → Add MCP → Add custom MCP. Older versions of the UI put this under Settings → MCP Marketplace → Add Your Own. 3. Choose Transport HTTP, set the Server URL to `https://mcp.finchreach.com/mcp`, choose Authentication Auth Header (the key defaults to `Authorization`), and set the value to `Bearer` followed by your key. 4. Save, then test the tools. Ask Devin to list your LinkedIn accounts to confirm the connection. 5. See the [Devin guide](https://finchreach.com/guides/devin) and the [quickstart](https://finchreach.com/docs) for more detail. Adding a server requires the “Manage MCP Servers” permission in Devin, and you can add it at organization or personal scope. ## Who in the org should be able to use the key? Think about this before you click save. A Finchreach API key is tied to one person’s LinkedIn account, here usually the founder’s. If you add it at organization scope in Devin, anyone who can start a Devin session in that org can have Devin read that person’s notifications and messages and queue posts, comments and messages in their name. - If only the founder should drive it, add the server at personal scope. - If the whole team should be able to triage, organization scope is fine, but agree that approvals stay with the account owner. The approval queue lives in the Finchreach dashboard, not in Devin. - Keep approval mode on. It is the default and it is the safeguard that makes shared use reasonable. - Revoke and recreate the key if someone with access leaves. ## Is it safe to let Devin act on a founder’s LinkedIn? Automating a personal LinkedIn account carries risk under LinkedIn’s User Agreement. The approval queue lowers that risk, because nothing goes live until a person approves it, but it doesn’t remove it. Keep the volume human: a handful of thoughtful replies, not a reply to every mention. There is also a quieter risk: text Devin reads on LinkedIn is written by strangers, and it may contain instructions aimed at the agent. Treat LinkedIn content as untrusted input, review issues before they are worked on, and read every queued comment before approving it. ## Questions people ask Can Devin reply directly under someone’s comment? No. Finchreach posts top-level comments on a post; there is no tool for threaded replies to a specific comment today. A direct message is the alternative for one-to-one follow-up. Does Devin need write access to LinkedIn for this? Only for the thank-you step. Finding reports and filing issues uses LinkedIn read tools plus Devin’s own access to your repo or tracker. Can it watch a company page instead of the founder’s account? No. Finchreach works with personal LinkedIn accounts, not company pages. Can it attach screenshots to a reply? No. Comments, posts and messages are text only. Where do I start? Set up the connection with the [Devin guide](https://finchreach.com/guides/devin), or read [what a LinkedIn MCP server is](https://finchreach.com/blog/what-is-a-linkedin-mcp-server) first.