gws-modelarmor-sanitize-prompt

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

Install with Tessl CLI

npx tessl i github:googleworkspace/cli --skill gws-modelarmor-sanitize-prompt

What are skills?

Review — 66%

Does it follow best practices?

If you maintain this skill, you can automatically optimize it using the tessl CLI to improve its score:

npx tessl skill review --optimize ./path/to/skill

Learn more

Validation — 8 / 11 Passed

Validation for skill structure

SKILL.md

Review

Evals

Evaluation results

100%

65%

LLM Gateway Prompt Safety Integration

Correct sanitize-prompt command and template format

Criteria

Without context

With context

Correct subcommand

100%

Not +sanitize-response

100%

--template flag present

100%

Full resource name format

100%

Project component correct

100%

Location component correct

100%

--text flag used

100%

Dynamic prompt argument

100%

Stdout not suppressed

100%

README present

100%

Without context: $0.1674 · 38s · 12 turns · 14 in / 2,138 out tokens

With context: $0.2224 · 47s · 16 turns · 54 in / 2,462 out tokens

100%

57%

Nightly Prompt Audit Pipeline

Stdin piping for batch prompt sanitization

Criteria

Without context

With context

Correct subcommand

100%

Not +sanitize-response

100%

--template flag present

100%

Full resource name format

53%

100%

Stdin pipe pattern

100%

No --text flag

33%

100%

No --json flag

100%

Iterates per line

100%

Results written to file

100%

Sample prompts.txt

100%

Without context: $0.1526 · 51s · 11 turns · 16 in / 2,678 out tokens

With context: $0.3551 · 1m 15s · 23 turns · 286 in / 3,967 out tokens

100%

55%

Advanced Model Armor Sanitization with Custom Request Body

JSON request body override

Criteria

Without context

With context

Correct subcommand

100%

Not +sanitize-response

100%

--template flag present

100%

Full resource name format

100%

--json flag used

100%

JSON contains prompt content

100%

No --json + --text combined

100%

No stdin pipe with --json

100%

Response saved to file

100%

demo_notes.md present

100%

Without context: $0.2964 · 1m 22s · 19 turns · 23 in / 4,437 out tokens

With context: $0.2045 · 44s · 14 turns · 19 in / 2,337 out tokens

Evaluated: about 11 hours ago
Agent: Claude Code

Table of Contents

LLM Gateway Prompt Safety Integration Nightly Prompt Audit Pipeline Advanced Model Armor Sanitization with Custom Request Body

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.