Back to skill

Security audit

angi-mcp

Security checks for vulnerabilities and agentic risk

Overview

The skill provides read-only Angi lookups and clearly separates optional access to your signed-in account.

Review the external CLI and browser extension before installing, since their code is outside this package. Keep access scoped to Angi and pair my.angi.com only when you want private account lookups. Pairing persists, and account requests expose retrieved information to your agent. The bundled helper was fully inspected, but external dependencies and live account behavior were not verified.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

Although the code uses the declared CLI and Angi domain, it reads private account resources rather than the described directory resources. Personal identity, project counts, and account review history are not disclosed by the declared capabilities to search professionals and read their profiles, ratings, and reviews.

Content

No source excerpt is available for this finding.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 79)May include surrounding context.

md
| node references/rsc.mjs legacyId --dedupe id \

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 83)May include surrounding context.

md
`rsc.mjs` exists because the data is **not** in the HTML and **not** `jq`-able.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 98)May include surrounding context.

md
`rsc.mjs` does *not* apply. Use `jq` directly:

The skill retrieves private signed-in account data beyond the manifest's directory-lookup scope.

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

The manifest describes finding service professionals, reading their profiles and reviews, and listing trade/city taxonomy. The commands at L101–L110 additionally use the browser's authenticated session to retrieve the user's identity, project data, and personally authored reviews from my.angi.com. Although documented as optional in the body, this private-account capability is not disclosed in the manifest description.

Content

No source excerpt is available for this finding.