entra-agent-id

40Medium

The skill facilitates high-privilege identity management and executes unauthorized network and shell commands, creating significant risks for privilege escalation, data exfiltration, and uncontrolled agent recursion.

Threat Analysis

AI Agent Traps ↗Scanned 6/30/2026
Content InjectionPerceptionclean
Semantic ManipulationReasoningclean
Cognitive StateMemory & Learningclean
SystemicMulti-Agent Dynamicsclean
Human-in-the-LoopHuman Overseerclean
HygienePackaging & Manifest5 findings

Skill Info

Namemicrosoft/azure-skills/entra-agent-id
Registrygithub
Version7e172d6883f1
PURLpkg:github/microsoft/azure-skills@7e172d6883f1?skill=entra-agent-id
Stars1,245
Installs142,500

Install

This skill has findings worth reviewing before installing.
GitHub
GitHub
142,500 installs
Skills.sh
Skills.shDocs
npx skills add https://github.com/microsoft/azure-skills

Assessments (3)

AI Agent Traps ↗
High-Privilege Identity Provisioningmediumdescription mismatchAML.T0011.000LLM06:2025LLM09:2025

The skill provides instructions for creating and managing high-privilege Entra identities and granting broad permissions (e.g., User.Read.All), which could be abused to escalate privileges or establish persistence if the agent is compromised. [ensemble: confirmed by 3/3 passes; severity set to the agreed median (ADR-0067).]

80% confidence
Granting per-Agent-Identity application (appRoleAssignments) or delegated (oauth2PermissionGrants) permissions
Runaway agent loop — excessive retry counts, deep recursion, or self-recursive call patternsmediumresource abuseAML.T0029AML.T0034LLM10:2025

Runaway agent loop — excessive retry counts, deep recursion, or self-recursive call patterns

95% confidenceline 322
retry 403
Direct shell/command execution sink in skill codelowcommand executionAML.T0050AML.T0053AML.T0072LLM05:2025LLM06:2025

Direct shell/command execution sink in skill code

95% confidenceline 4
subprocess.run(
    ["az", "ad", "signed-in-user", "show", "--query", "id", "-o", "tsv"],
    capture_output=True, text=True, check=True,
)
Outbound network request primitive in skill code — can transmit data externallylowdata exfiltrationAML.T0025AML.T0024LLM02:2025

Outbound network request primitive in skill code — can transmit data externally

95% confidenceline 15
requests.post(
    f"{GRAPH}/applications/microsoft.graph.agentIdentityBlueprint",
    headers=headers, json=blueprint_body,
)
Outbound network request primitive in skill code — can transmit data externallylowdata exfiltrationAML.T0025AML.T0024LLM02:2025

Outbound network request primitive in skill code — can transmit data externally

95% confidenceline 2
requests.post(
    f"{GRAPH}/servicePrincipals/microsoft.graph.agentIdentityBlueprintPrincipal",
    headers=headers, json=sp_body,
)
Outbound network request primitive in skill code — can transmit data externallylowdata exfiltrationAML.T0025AML.T0024LLM02:2025

Outbound network request primitive in skill code — can transmit data externally

95% confidenceline 8
requests.post(
    f"{GRAPH}/servicePrincipals/microsoft.graph.agentIdentity",
    headers=headers, json=agent_body,
)
Outbound network request primitive in skill code — can transmit data externallylowdata exfiltrationAML.T0025AML.T0024LLM02:2025

Outbound network request primitive in skill code — can transmit data externally (seen 2 times in this file at lines 1, 8)

95% confidenceline 1
requests.get(
    f"{GRAPH}/servicePrincipals?$filter=appId eq '00000003-0000-0000-c000-000000000000'",
    headers=headers,
)
Outbound network request primitive in skill code — can transmit data externallylowdata exfiltrationAML.T0025AML.T0024LLM02:2025

Outbound network request primitive in skill code — can transmit data externally

95% confidenceline 5
requests.post(
    f"{GRAPH}/oauth2PermissionGrants",
    headers=headers,
    json={
        "clientId": agent_sp_id,
        "consentType": "AllPrincipals",
        "resourceId": graph_sp["id"],
        "scope": "User.Read Tasks.ReadWrite Mail.Send",
        "expiryTime": expiry,
    },
)
Missing allowed-tools declarationlowunauthorized tool useAML.T0053LLM06:2025

Skill executes commands, writes files, or accesses the network but declares no allowed-tools, so its tool surface cannot be reviewed or constrained.

90% confidence
Undeclared network usagelowunauthorized tool useAML.T0053LLM06:2025

Skill contains network access patterns (HTTP requests, socket connections) without declaring network capability in allowed-tools.

60% confidence
Referenced file missing from skill packagelowpolicy violation

SKILL.md links to "references/oauth2-token-flow.md" but the file is not part of the skill package — the workflow silently degrades or the content is sourced elsewhere at runtime

90% confidence
[references/oauth2-token-flow.md](references/oauth2-token-flow.md)
Referenced file missing from skill packagelowpolicy violation

SKILL.md links to "references/obo-blueprint-setup.md" but the file is not part of the skill package — the workflow silently degrades or the content is sourced elsewhere at runtime

90% confidence
[references/obo-blueprint-setup.md](references/obo-blueprint-setup.md)
Referenced file missing from skill packagelowpolicy violation

SKILL.md links to "references/runtime-token-exchange.md" but the file is not part of the skill package — the workflow silently degrades or the content is sourced elsewhere at runtime

90% confidence
[references/runtime-token-exchange.md](references/runtime-token-exchange.md)
Referenced file missing from skill packagelowpolicy violation

SKILL.md links to "references/sdk-sidecar-deployment.md" but the file is not part of the skill package — the workflow silently degrades or the content is sourced elsewhere at runtime

90% confidence
[references/sdk-sidecar-deployment.md](references/sdk-sidecar-deployment.md)
Referenced file missing from skill packagelowpolicy violation

SKILL.md links to "references/sdk-sidecar.md" but the file is not part of the skill package — the workflow silently degrades or the content is sourced elsewhere at runtime

90% confidence
[references/sdk-sidecar.md](references/sdk-sidecar.md)

Badge

Skill Check badge
Markdown
[![Mondoo Skill Check](https://mondoo.com/ai-agent-security/api/badge/github/microsoft/azure-skills/entra-agent-id.svg)](https://mondoo.com/ai-agent-security/skills/github/microsoft/azure-skills/entra-agent-id)
HTML
<a href="https://mondoo.com/ai-agent-security/skills/github/microsoft/azure-skills/entra-agent-id"><img src="https://mondoo.com/ai-agent-security/api/badge/github/microsoft/azure-skills/entra-agent-id.svg" alt="Mondoo Skill Check" /></a>
Image URL
https://mondoo.com/ai-agent-security/api/badge/github/microsoft/azure-skills/entra-agent-id.svg

Secure your AI agents

Skills can read files, run commands, and access credentials. Mondoo helps organizations manage the security risks of AI agent skills across their entire fleet, discover and control Shadow AI usage.

  • Continuous scanning of skills across all registries
  • Policy enforcement before skills reach your agents
  • Integration with your existing security workflow

By submitting, you agree to our privacy policy.