Model Armor Prompt Sanitizer
Screen user prompts through a Google Model Armor template before sending them to a model.
- Repository
- googleworkspace/cli
- Stars
- 32,059
- Licence
- Apache-2.0
- Files
- 1 file
What it does
This skill teaches an agent to run the gws command-line tool so that a user's text is checked against a Google Cloud Model Armor template, a service that screens content for safety issues such as prompt injection or harmful material. The agent can pass text directly, a full JSON request, or piped stdin, and receives the sanitization result. It is aimed at checking inbound user input before an AI model processes it.
Great for
- Developers building AI apps on Google Cloud who want to screen user prompts before they reach a model
- Security or platform teams adding a safety check step to an existing Model Armor setup
- Agents that need to verify user input against a configured Model Armor template
Skip it if
- Checking the model's outgoing responses; the skill's own tips say to use +sanitize-response for that direction
- Setups without a Google Cloud project and an existing Model Armor template, since the required --template resource name cannot be supplied
- Treating it as a complete safety guarantee; it only applies the filters configured in the chosen template
Try asking
- “Sanitize this user prompt against my Model Armor template projects/my-proj/locations/us-central1/templates/strict before I send it to the model”
- “Run this incoming chat message through Model Armor and tell me whether it was flagged”
- “Pipe this support ticket text through my prompt sanitization template and report the result”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.
gws-modelarmor-sanitize-prompt.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip gws-modelarmor-sanitize-prompt.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip gws-modelarmor-sanitize-prompt.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Provide the full template resource name in the form projects/PROJECT/locations/LOCATION/templates/TEMPLATE.
- Use --json for a full request body when you need settings beyond plain text; it overrides --text.
- Run gws generate-skills first if the shared gws-shared skill file is missing.
What's inside
- SKILL.md
Redistributed unmodified from googleworkspace/cli under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.