Back to skill

Security audit

SocialKit YouTube Pipeline

Security checks for vulnerabilities and agentic risk

Overview

The skill provides a user-approved YouTube workflow, with a progress-reporting bug that does not bypass its required verification command.

Review the separate companion skills before granting YouTube access. Personally approve each checkpoint and require successful --verify results before continuing; --status can display stale approvals when project files are damaged or missing. The status bug should be fixed, and approval records should be treated as an audit trail rather than tamper-proof authorization.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (8)

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 27)May include surrounding context.

md
`node scripts/state.mjs --init --dir yt-<slug> --slug <slug>`.

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 32)May include surrounding context.

md
terminal: `node scripts/state.mjs --dir yt-<slug> --approve script`.

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 33)May include surrounding context.

md
Then run `node scripts/state.mjs --dir yt-<slug> --verify script`. Continue only

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 40)May include surrounding context.

md
user to run `node scripts/state.mjs --dir yt-<slug> --approve metadata`, then verify.

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 44)May include surrounding context.

md
`node scripts/state.mjs --dir yt-<slug> --approve upload`, then verify.

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 47)May include surrounding context.

md
Check progress at any time with `node scripts/state.mjs --dir yt-<slug> --status`.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: env.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Status reports an existing approval as valid when its covered files cannot be determined, contradicting the documented invalidation guarantee.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The status path silently catches errors while determining checkpoint files and then verifies an empty file list. If an approval record exists, verification returns success without checking any hashes, so deleting or corrupting metadata.json or upload-plan.json can make status falsely report an approval as valid. This undermines the workflow's mandatory human-review guarantees, although the separate --verify path does not swallow the same errors.

Content

No source excerpt is available for this finding.