$ loading_
Detect leaked secrets, prompt injection, and PII in AI output.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "io.github.spektre-labs/sigma-gate" MCP server from askskill: Run: claude mcp add --transport http 'io-github-spektre-labs-sigma-gate' 'https://sigma-gate-864996675261.us-central1.run.app/mcp'
Check this AI output for leaked secrets, prompt injection, or personal data, and return matches plus a risk verdict: {{AI output content}}A safety screening result showing whether the content contains secret leakage, injection, or PII risks.
Run a trust-gate check on the following generated content before publishing; if secrets, prompt injection, or PII are found, mark it as blocked: {{text to inspect}}A decision-oriented result for pre-release checks, such as pass or block with reasons.
Run deterministic safety checks on this batch of AI output samples for leaked secrets, prompt injection, and personal data, then summarize the risk counts: {{sample list}}Per-sample findings and an overall risk summary for evaluating output quality and safety.
Developers or product managers can add a deterministic check after AI generation to detect leaked secrets, prompt injection, and personal data before content is exposed.
In chatbots, text generation, or automated workflows, this tool can serve as a trust gate to apply consistent safety screening to model outputs.
Teams can use it to inspect model samples for sensitive leaks, injection content, or personal data as part of output safety testing and quality control.
It is a deterministic trust gate for AI output that checks for leaked secrets, prompt injection, and personal data risks in one call.
Based on the description, it focuses on three risk types: leaked secrets, prompt injection, and PII. For more detailed rules or output format, see the source repository.
The provided information does not include installation steps, runtime requirements, or key prerequisites. Please see the source repository for details.
Audit AI agent code for pre-deploy security risks missed by generic SAST.
Add deterministic security checks, host audits, and diagnostics to AI workflows.
Score agent outputs with guardrails, policy checks, injection, and PII detection.
Add human approval and tamper-evident logs to risky AI agent actions.
Scan prompts locally and redact sensitive data before sending to LLMs.
Scan inbound and outbound emails for prompt injection, phishing, and secret leaks.