deploy-model

SAFE max severity MEDIUMSkillSpector Caution 28 registry microsoft last scanned 2026-10-22 2 scans

Verdict safe, but 2 findings (1 medium, 1 low), mainly data exfiltration, unauthorized tool use. Most recent scan 2026-10-22.

Findings2

Critical0

High0

Medium1

Low1

Analyzersstatic bytecode pipeline correlation behavioral trigger

Scanner version2.3.1

Second scanner: NVIDIA SkillSpector

VerdictCaution 28

Score28 / 100

Issues2

CategoriesAgent Snooping Excessive Agency

Version2.12.0 (static)

Independent static run, no language model, over the same files as the Cisco scan above. How the two scanners compare across the whole corpus: Two scanners, one corpus.

What this skill does

Unified Azure OpenAI model deployment skill with intelligent intent-based routing. Handles quick preset deployments, fully customized deployments (version/SKU/capacity/RAI policy), and capacity discovery across regions and projects. USE FOR: deploy model, deploy gpt, create deployment, model deployment, deploy openai model, set up model, provision model, find capacity, check model availability, where can I deploy, best region for model, capacity analysis. DO NOT USE FOR: listing existing deployments (use foundry_models_deployments_list MCP tool), deleting deployments, agent creation (use…

Package: 17 files, 102 KB
SKILL.md: 905 words, 11 sections
Code: PowerShell (3), Shell (3)
License: MIT

SKILL.md covers: Deploy Model · Quick Reference · Intent Detection · Routing Rules · Multi-Mode Chaining · Project Selection (All Modes) · Resolution Order · Confirmation Step (Required) · Pre-Deployment Validation (All Modes) · Prerequisites · Sub-Skills

Findings

MEDIUM Bash: tainted data flows to dangerous sink Data Exfiltrationbehavioral

Variable $RESOURCE_GROUP (line 5) flows to `source` at line 17.

preset/references/preset-workflow.md:70

echo "  Resource Group: $RESOURCE_GROUP"

Remediation: Review the data flow from source to sink. Avoid sending sensitive or untrusted data to network or execution commands.

LOW Script contacts a destination the documentation never mentions Unauthorized Tool Usestatic

Script contacts or configures 'ai.azure.com' but the skill documentation never visibly mentions it. Undocumented network destinations can indicate behaviour the manifest and description do not disclose.

scripts/generate_deployment_url.sh:87

Remediation: Document the destination in SKILL.md (or the manifest description) if it is intended, or remove the undisclosed network access. Organizations can list internal hosts in llm_analysis.trusted_reference_domains.

What is in the package

File	Kind	Size
`SKILL.md`	doc	7.1 KB
`TEST_PROMPTS.md`	doc	3.3 KB
`capacity/SKILL.md`	doc	6.8 KB
`customize/EXAMPLES.md`	doc	4.3 KB
`customize/SKILL.md`	doc	8.8 KB
`preset/EXAMPLES.md`	doc	2.9 KB
`preset/SKILL.md`	doc	4.8 KB
`scripts/generate_deployment_url.ps1`	script	2.3 KB
`scripts/generate_deployment_url.sh`	script	2.6 KB
`capacity/scripts/discover_and_rank.ps1`	script	4.6 KB
`capacity/scripts/discover_and_rank.sh`	script	4.5 KB
`capacity/scripts/query_capacity.ps1`	script	3.0 KB
`capacity/scripts/query_capacity.sh`	script	2.9 KB
`customize/references/customize-guides.md`	doc	3.4 KB
`customize/references/customize-workflow.md`	doc	13 KB
`preset/references/preset-workflow.md`	doc	22 KB
`preset/references/workflow.md`	doc	5.6 KB

External hosts referenced

management.azure.com ai.azure.com

Hostnames that appear in the skill's files. Referencing a host is not a finding; compare against the findings above and the skill's stated purpose.

Environment variables referenced

MODEL_NAME ACCOUNT_NAME PROJECT_RESOURCE_ID SUBSCRIPTION_ID MODEL_VERSION PROJECT_NAME MODEL_FORMAT SELECTED_REGION PROJECT_REGION SUB_ID ALL_REGIONS_JSON NEW_PROJECT_NAME TENANT_INFO ORG_NAME REGION AVAILABLE_REGIONS PROJECTS_IN_REGION REGIONS_WITH_CAP BASE_URL DEPLOYMENT_PATH DEPLOYMENT_URL ENCODED_PATH ENDPOINT PROJECTS_JSON PROJECT_COUNT REGIONS UNAVAILABLE_REGIONS

Package contents recorded 2026-10-22 at commit f49ac319a7d3. The scan above ran against the version current at scan time; the two can differ if the skill has since been updated.

What we checked

Static analyzer. Pattern rules over SKILL.md and bundled files: known-bad commands, exfiltration URLs, secrets, obfuscation.
Bytecode analyzer.
Pipeline analyzer.
Correlation analyzer.
Behavioral analyzer. Checks what the skill asks the agent to do against what its description claims it does.
Trigger analyzer. Looks for instructions that fire on conditions the user did not ask for.

What we did not check

Runtime behaviour. The skill was not executed; findings come from reading its files.
Anything fetched at install or run time, including remote scripts, package dependencies, and URLs the skill tells the agent to open.
Content the analyzers cannot read: images, binaries, and some non-English text.
Known scanner bypass classes. See the methodology and limitations page.

Scan history

Date	Verdict	Findings	Change	Analyzers	Raw
2026-10-22	SAFE	2	=	static bytecode pipeline correlation behavioral trigger	JSON
2026-10-22	SAFE	2	first scan	static bytecode pipeline correlation behavioral trigger	JSON

Sources

Repository: https://github.com/microsoft/azure-skills
Raw scan result: microsoft-azure-skills-skills-microsoft-foundry-models-deploy-model-scan.json
Badge for your README: ![skillscan.io](https://skillscan.io/badge/microsoft-deploy-model.svg)
Request a re-scan