AI agents use like-post to create or update resources in Bluesky — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Bluesky environment.
Liking a post creates a new like record, which is a write operation that modifies data (the post's like count/list) reversibly. Unlike posts can be unliked later via the 'unlike-post' sibling tool, confirming reversibility. While this modifies state on a social platform, the blast radius is minimal—it has no destructive, financial, or external code execution consequences.
From the tool's definition Tool name 'like-post' and description 'Like a post' indicate a reversible modification action that creates a like record on a social media post.
Documented attack patterns abuse exactly the kind of access like-post gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Bluesky, and nothing reaches the server without passing your rules. This is the rule we recommend for like-post:
{
"version": "1",
"default": "deny",
"tools": {
"like-post": {
"limits": [
{
"counter": "like-post_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}like-post stays usable, but capped — an agent stuck in a loop can't make hundreds of changes a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Like a post. It is categorised as a Write tool in the Bluesky MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Bluesky MCP server in PolicyLayer and add a rule for like-post: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Bluesky. Nothing to install.
like-post is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the like-post rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for like-post. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
like-post is provided by the Bluesky MCP server (semioz/bluesky-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Bluesky, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
11 Bluesky tools catalogued and risk-classified — across an index of 43,000+ MCP servers.