aidefence_scan
Scan input text for AI manipulation threats (prompt injection, jailbreaks, PII). Returns threat assessment with <10ms latency. Use when nothing native exists — Claude Code does not have a PII / prompt-injection / adversarial-text scanner. Pair with any tool that ingests untrusted input (browser s...
This record as markdown: /tools/io-github-ruvnet-claude-flow/aidefence-scan.md
What aidefence_scan does on Claude Flow
AI agents call aidefence_scan to retrieve information from Claude Flow without modifying anything. It is typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Why aidefence_scan is rated Low
This is a defensive security scanner that classifies and detects threats in text without causing side effects. It retrieves a threat assessment from input analysis, matching the Read category (query/analyze without modification).
From the tool's definition Scan input text for AI manipulation threats (prompt injection, jailbreaks, PII). Returns threat assessment — the tool analyzes and returns a report without modifying, deleting, or executing code based on input.
Attacks that exploit this kind of access
The rule that runs aidefence_scan safely
PolicyLayer is an MCP gateway: it sits between your AI agents and Claude Flow, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For aidefence_scan, this is the rule to start with:
aidefence_scan is read-only, so it stays allowed. Everything else on the server is denied unless you say otherwise.
The button opens the PolicyLayer dashboard: create your workspace, connect Claude Flow, apply this rule, and every aidefence_scan call is checked against it from then on.
Questions about aidefence_scan
Scan input text for AI manipulation threats (prompt injection, jailbreaks, PII). Returns threat assessment with <10ms latency. Use when nothing native exists — Claude Code does not have a PII / prompt-injection / adversarial-text scanner. Pair with any tool that ingests untrusted input (browser scrape, federation envelope, memory_import_claude). It is categorised as a Read tool in the Claude Flow MCP Server, which means it retrieves data without modifying state.
Register the Claude Flow MCP server in PolicyLayer and add a rule for aidefence_scan: 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 Claude Flow. Nothing to install.
aidefence_scan is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the aidefence_scan 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 aidefence_scan. 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.
aidefence_scan is provided by the Claude Flow MCP server (claude-flow). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
More on Claude Flow, and thousands of servers like it.
This server
Across the catalogue