ANTHROPIC INTEGRATION
Bloom Security integrates with the Anthropic Compliance API to give security teams full visibility into how Claude is used across your organization, from generated code and uploaded files to new plugins, skills, MCP servers, and configuration changes.
Developers generate scripts, share files, install plugins, and register MCP servers. Anthropic’s Compliance API provides the raw event data, but events alone don’t tell you what’s risky. That’s where Bloom comes in.
Is anyone sharing production secrets with Claude?
What plugins and MCP servers have been introduced?
Did someone disable safety checks?
What code is Claude generating and is it safe?
CAPABILITIES
01
Bloom retrieves every artifact Claude generates and every file users upload. It scans for hardcoded credentials, API keys, connection strings, PII, and sensitive configurations—before they become incidents.
02
Every new plugin, skill, command, and MCP server is a new component in your AI supply chain. Bloom detects when they’re introduced, tracks whether they’ve been reviewed, and flags unscanned components.
03
When someone changes Claude’s execution policy, like disabling permission prompts, Bloom detects it immediately and alerts your security team. One click to deploy a guardrail.
04
All Claude activity across every endpoint, user, and session in one place. Filter by endpoint, session, or user. See what happened, when, and whether it’s safe.
HOW IT WORKS
1
Connect
Paste your Anthropic Compliance Access Key in Bloom Settings.
2
Ingest
Bloom polls the Activity Feed, retrieving events.
3
Scan
Artifacts, files, plugins, and MCP servers are analyzed for security risks.
4
Act
View analyzer results in the dashboard. Deploy guardrails with one click.
COVERAGE
Visibility and control for the AI tools your team uses every day.