Gws Modelarmor Sanitize Prompt

Google Model Armor: Sanitize a user prompt through a Model Armor template.

Security Scan

The skill is an instruction-only wrapper around the gws CLI to run a Model Armor prompt-sanitization template and its requirements and instructions are generally consistent with that purpose.

Purpose & Capability

Name/description match the runtime instructions: the SKILL.md simply invokes 'gws modelarmor +sanitize-prompt' and requires the 'gws' binary. Requiring the gws CLI is proportionate for a wrapper around Google Model Armor functionality.

Instruction Scope

Instructions are narrow and only describe calling the gws CLI and reading stdin/flags. They refer to a sibling file '../gws-shared/SKILL.md' for auth/global flags/security rules; that external dependency is expected but means behavior (auth, where data is sent, logging) depends on that other SKILL.md which is not included here.

Install Mechanism

No install spec and no code files — lowest risk. The skill relies on an existing 'gws' binary; nothing is downloaded or written by the skill itself.

Credentials

The skill declares no env vars or credentials itself, which is reasonable because the gws CLI likely handles auth. Because auth is delegated to '../gws-shared/SKILL.md' or the gws environment, you should review that shared config to confirm which credentials will be used and whether they are appropriate for prompt sanitization.

Persistence & Privilege

The skill does not request always:true and is user-invocable. It does not modify other skills or request system-wide persistence.

Assessment

This is an instruction-only skill that just calls the external 'gws' CLI to sanitize prompts. Before installing, verify you trust the 'gws' binary on your system and inspect ../gws-shared/SKILL.md (or your gws auth configuration) to see which credentials and endpoints will be used and whether any automatic config generation (gws generate-skills) would write files you don't expect. Avoid sending highly sensitive secrets to the sanitizer unless you confirm the Model Armor template and gws configuration handle data the way you intend.

modelarmor +sanitize-prompt

PREREQUISITE: Read ../gws-shared/SKILL.md for auth, global flags, and security rules. If missing, run gws generate-skills to create it.

Sanitize a user prompt through a Model Armor template

Usage

gws modelarmor +sanitize-prompt --template <NAME>

Flags

Flag	Required	Default	Description
--template	✓	—	Full template resource name (projects/PROJECT/locations/LOCATION/templates/TEMPLATE)
--text	—	—	Text content to sanitize
--json	—	—	Full JSON request body (overrides --text)

Examples

gws modelarmor +sanitize-prompt --template projects/P/locations/L/templates/T --text 'user input'
echo 'prompt' | gws modelarmor +sanitize-prompt --template ...

Tips

• If neither --text nor --json is given, reads from stdin.
• For outbound safety, use +sanitize-response instead.

See Also

• gws-shared — Global flags and auth
• gws-modelarmor — All filter user-generated content for safety commands