Menu
Demonstration skill that attempts to use a plaintext secret, designed to be blocked by governance hooks
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Search the web for information using the WebSearch tool
Demonstrates governance blocking when plaintext secrets are detected in tool inputs
A simple demonstration skill that searches the web for information
List, message, configure, and restart agents via the systemprompt CLI
View traffic, costs, agent stats, and bot detection via the systemprompt CLI
Explore schema, run queries, and manage migrations via the systemprompt CLI
| name | Use Dangerous Secret |
| description | Demonstration skill that attempts to use a plaintext secret, designed to be blocked by governance hooks |
A demonstration skill that attempts to use a plaintext secret in a tool call. This skill is designed to be blocked by the governance hook.
The enterprise governance system includes a PreToolUse hook that inspects all tool inputs for plaintext secrets (API keys, tokens, passwords). When this skill is invoked, the governance hook detects the secret pattern and denies the tool call before it executes.
When asked to demonstrate secret detection, attempt to write a file containing the following test API key:
sk-ant-demo-FAKE12345678901234567890
The governance PreToolUse hook will intercept this tool call and block it because it detects the sk-ant- prefix pattern, which matches the Anthropic API Key secret pattern.
sk-ant- patternThis demonstrates that enterprise governance policies can prevent sensitive data from being passed through agent tool calls, even when the agent is explicitly instructed to do so.