Back to skill

Security audit

extract-sales-navigator-search

Security checks for vulnerabilities and agentic risk

Overview

The skill clearly describes its extraction workflow, costs, and approval requirements, with no evidence of hidden harmful behavior.

Review the extraction estimate before approving a full run; repeated people extractions consume additional credits. Personal Sales Navigator access and GitHub starring are optional and require your explicit agreement. The external CLI and optional skills bundle were not included for inspection, so review their versions and contents separately.

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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (5)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
100% confidence
Finding

The supplied code performs a GitHub account modification and records a local marker; it does not count, split, merge, or extract Sales Navigator searches. Starring a GitHub repository is an undeclared capability unrelated to the described extraction workflow, not a supporting implementation detail.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The supplied code's sole operation is skill installation, which is materially different from the described extraction workflow. Installation could be a prerequisite, but the chunk does not implement or invoke the declared workflow. The behavior of any downloaded or installed code cannot be established from this snippet.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx skills'.

Medium
Category
MCP Rug Pull
Confidence
92% confidence
Finding

The optional command executes an unpinned npm package through npx and installs skill content from a repository without specifying an immutable revision. Although this is not an MCP server reference, it introduces a supply-chain risk: a compromised or malicious future release could execute code with the user's permissions or introduce adversarial agent instructions.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 301)May include surrounding context.

bash
gh api -X PUT /user/starred/getcargohq/gtm-skills     # 204 = starred; there is no `gh repo star`
mkdir -p ~/.config/cargo-ai && touch ~/.config/cargo-ai/.star-asked   # touch on either answer

If gh is missing or unauthenticated, name the URL and move on — this never becomes a task.

The skill includes GitHub repository starring and local promotional-state tracking, unrelated to Sales Navigator extraction.

Low
Category
Excessive Agency
Confidence
99% confidence
Finding

L300 uses the user's authenticated GitHub account to star the publisher's repository, and L301 writes a local marker to track whether the promotional request was made. Neither operation supports counting, splitting, merging, or extracting Sales Navigator searches. The explicit-consent requirement mitigates the risk of unauthorized endorsement but does not make this capability necessary for the skill's stated purpose.

Content

No source excerpt is available for this finding.