sketch-security-guardrails

Name: sketch-security-guardrails
Availability: InStock
Author: klabo

Enforce removal or hashing of API keys and raw images in logs.

Updated Dec 31, 2025

Search Tags:#debugging #security #logs #secrets #telemetry #privacy #image-handling

Ready to run

npx skills add https://github.com/joelklabo/markdowntown-cli --skill sketch-security-guardrails

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: sketch-security-guardrails
Source: https://github.com/joelklabo/markdowntown-cli/tree/main/codex/skills/sketch-security-guardrails
Command: npx skills add https://github.com/joelklabo/markdowntown-cli --skill sketch-security-guardrails

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Prevent accidental exposure of API keys, bearer tokens, raw images, full prompts, and other sensitive data when debugging, capturing proof videos, or collecting telemetry for image-focused AI tools.

Core Features & Use Cases

Never log secrets: guidance to avoid logging API keys, auth headers, or bearer tokens.
Never log raw images: rules to prohibit base64 or binary image payloads in logs and error reports.
Minimal telemetry & safe debugging: recommendations to log only metadata (duration, provider, error codes), use hashed or truncated prompt representations, and use sample files or stubbed responses for demos and proof videos.
Review and scrub before sharing: checklist items to scrub screenshots, console output, and recorded sessions to remove PII and secrets.

Quick Start

Use these guardrails to audit and remove any API keys, raw image payloads, or full prompt content from logs and telemetry before sharing or publishing debug traces.

Frequently Asked Questions about sketch-security-guardrails

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I prevent API keys and bearer tokens from leaking in debugging logs?▼

Prevent API keys and bearer tokens from leaking in debugging logs by enforcing guardrails that remove or hash secrets before any telemetry or error trace is generated. This ensures sensitive authentication data is never captured in log outputs.

What is the best way to stop raw image bytes from appearing in error reports?▼

Stop raw image bytes from appearing in error reports by applying strict logging rules that prohibit base64 or binary image payloads in logs and error traces. Log only metadata like duration, provider, and error codes instead of image data.

How do I safely truncate or hash full prompts in AI-assisted design telemetry?▼

Safely truncate or hash full prompts in AI-assisted design telemetry by configuring your logging pipeline to apply prompt hashing or truncation rules. This prevents sensitive prompt content from being fully exposed in debug traces or proof videos.

Can I use stubbed responses and sample files for safe proof-video workflows?▼

Yes, you can use stubbed responses and sample files for safe proof-video workflows to avoid capturing real sensitive data. This approach demonstrates functionality while keeping raw images and live API keys out of recorded sessions.

What should I check before sharing console output and screenshots from image-editing tools?▼

Before sharing console output and screenshots from image-editing tools, perform a review and scrub checklist to remove PII, secrets, and raw image payloads. This ensures no sensitive data accidentally leaks in shared debug traces.