Field Manual Action matrix
Platform action matrix
One table, every platform. What Blowhorn can do, and how each platform's own rules grade it. Read the notes before you charge.
Sources checked 13 October 2026.
Every action, every platform
What Blowhorn does on each platform, and how risky each action is by that platform's own rules. Scroll sideways on a phone.
| Platform | Post | Comment or reply | Amplify | Follow | Connect or invite | Accept invitations | Message |
|---|---|---|---|---|---|---|---|
| LinkedInYour Chrome, through the extension | Use careNo bait, no recycled posts [1][2] | RiskyLinkedIn bans automated comments [3] | RiskyRepost plus like from many profiles looks like a pod [3][4] | Not supported | Use careInvitation limits; restrictions last about a week [5] | SafeFlagged invitations are always skipped [6] | Not supported |
| XYour Chrome, through the extension | Use careNever the same text from two accounts [7] | SafeReplies are what the ranking rewards most [8] | RiskySeveral owned accounts reposting one post is named in X's rules [9] | Use careCapped at 400 a day, 15 per 15 minutes [10] | Not supported | Not supported | Not supported |
| RedditYour Chrome, through the extension | Use careMostly not promotion; subreddit rules first [11] | Use careNever the same comment twice [11] | RiskyUpvotes from several accounts are vote manipulation [12] | Not supported | Not supported | Not supported | Not supported |
| Hacker NewsYour Chrome, through the extension | Use care2 link submissions a day at most, never edited [10][13] | Use careYour own words only [13] | Not supportedBlowhorn never votes on Hacker News | Not supported | Not supported | Not supported | Not supported |
| BlueskyOfficial API, app password | SafeSpam and repetition are the limits [14] | Safe | Use careNo manipulated social signals [14] | Use careSame rule; keep it human-sized [14] | Not supported | Not supported | Not supported |
| GitHubOfficial API, your token | Not supported | Not supported | RiskyAll five reactions from many profiles; GitHub bans inauthentic interactions [15] | RiskyGitHub names automated following as rank abuse [15] | Not supported | Not supported | Not supported |
| SlackOfficial API, your token | Not supported | Not supported | Not supported | Not supported | Not supported | Not supported | Use careEach workspace sets its own rules. Hypothesis: one message per channel |
| Facebook, YouTube, Instagram, TikTok, Threads, MastodonNot supported by Blowhorn | Not supported | Not supported | Not supported | Not supported | Not supported | Not supported | Not supported |
- Safe: the platform allows it when a real person means it
- Use care: allowed, with limits or wording the platform spells out
- Risky: the platform's rules name it, or something close to it, as a reason to limit reach or restrict an account
How to read the grades
- Safe means the platform allows the action when a real person means it. It never means guaranteed: automation can still be blocked, and sessions can still break (Platforms).
- Use care means it's allowed with limits, or with wording rules the platform spells out. Read the note.
- Risky means the platform's own rules name this behavior, or something close to it, as a reason to cut reach or restrict an account. Use it rarely, from one profile, or not at all.
- Not supported means Blowhorn doesn't do it there. A blank cell means the same.
A grade describes the action, not the tool. LinkedIn says it doesn't permit "browser extensions that... automate activity on LinkedIn's website" (LinkedIn Help). Blowhorn types at human pace and hands you every CAPTCHA, but it is a browser extension acting on LinkedIn, so every LinkedIn action carries that rule with it. Weigh that before you run anything there unattended.
Notes and sources
- LinkedIn: "We're improving our systems to reduce repetitive, low-substance posts and engagement bait, such as 'comment to agree' prompts" (LinkedIn Pressroom, 12 Mar 2026).
- LinkedIn: "Don't do things to artificially increase engagement with your content" (Professional Community Policies).
- LinkedIn: automated comments "aren't allowed on LinkedIn"; an engagement pod "is not allowed on LinkedIn" (LinkedIn Pressroom, 13 Mar 2026).
- LinkedIn: "don't agree with others ahead of time to like or re-share each other's content" (Professional Community Policies).
- LinkedIn: "This restriction typically lasts one week" (Invitation limit reached).
- Why flagged invitations are skipped.
- X: "You may not post duplicative or substantially similar posts on one account or over multiple accounts you operate" (X automation rules).
- X's published 2023 weights: reply 13.5, like 0.5. Weights scale predicted probabilities, not raw counts (twitter/the-algorithm-ml).
- X: "operating multiple accounts that interact with the same or substantially similar content... in order to inflate or manipulate the prominence of content" (X platform manipulation policy).
- Blowhorn's built-in limits (Platforms). These are Blowhorn's caps, not numbers the platform publishes.
- Reddit: spam includes "Repeatedly posting the same or similar comments in a thread, community or across communities"; some communities use a 10% self-promotion rule (Reddit Help).
- Reddit prohibits "Creating and employing multiple accounts, voting services, or any automation to manipulate vote counts" (Reddit Help).
- Hacker News: "Don't solicit upvotes, comments, or submissions" and "Don't post generated text or AI-edited text" (HN guidelines).
- Bluesky: "Do not artificially manipulate features or social signals to gain unearned reach" (Community Guidelines, 19 Sep 2025).
- GitHub does not allow "inauthentic interactions, such as fake accounts and automated inauthentic activity; rank abuse, such as automated starring or following" (GitHub Acceptable Use Policies).
More from the Field Manual
Platform briefing
LinkedIn
What LinkedIn's feed rewards (time spent reading, real conversation, relevance to your field) and what it cuts (pods, automated comments, bait, recycled posts), from LinkedIn's own engineering and policy pages.
Platform briefing
X
What X's For You feed rewards (replies, conversations, time spent, shares) and what sinks a post (mutes, blocks, reports, duplicates, coordinated accounts), straight from the ranking code X published.
Platform briefing
Reddit
What Reddit rewards (being a real member of a community, early votes, fresh posts) and what gets you removed or banned (spam, self-promotion overload, repeated comments, vote manipulation), from Reddit's own rules and code.
Platform briefing
Hacker News
What Hacker News asks of posters (curiosity first, your own words, no solicited votes) and why it's the least forgiving front, from its own guidelines.
Platform briefing
Bluesky
How Bluesky's feeds work (your Following feed, Discover, and custom feeds anyone can build) and the one rule that matters most, no manipulated signals.
General education
Other feeds
What the other big feeds say they reward and reduce, in their own words. General education only. Blowhorn doesn't post to these platforms.
For CLI and agent users
CLI tuning
Prescriptive settings for the Blowhorn CLI and blowhorn-axi that map each platform rule in this manual to a flag or default. For operators and the agents they run.