Back to skill

Security audit

codex-delegation

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marked this skill suspicious because it reports that file contents may be sent to OpenAI without guidance on excluding sensitive data; the scan was partial.

Review the skill’s data-sharing implications before installing. Consider whether files from your projects could contain credentials or other sensitive information, and whether you are comfortable with selected contents being sent to OpenAI. The scan was partial because two local path references could not be resolved, so the evidence is incomplete.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Env Variable Harvesting

High
Category
Data Exfiltration
Confidence
70% confidence
Finding

Code enumerates, copies, or searches environment variables for secrets. Bulk environment access can collect credentials unrelated to the skill's stated purpose.

Content

Scanner excerpt · SKILL.md (reported line 57)May include surrounding context.

Either persist an API key or log in interactively:

bash
printenv OPENAI_API_KEY | codex login --with-api-key

Verification runs codex login status, which reports one of "Logged

External Model or Provider Selection

High
Category
Excessive Agency
Confidence
90% confidence
Finding

Skill selects an external model or provider that may use a different account or billing plan than the operator expects. Undisclosed model switches can cause unexpected cost or quota consumption.

Content

Scanner excerpt · SKILL.md (reported line 86)May include surrounding context.

Direct CLI usage

bash
codex exec "Explain the retry logic in this package"

Smart Delegation

The workflow includes file contents in prompts sent to OpenAI without warning users about disclosure of potentially sensitive source data.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The workflow explicitly requires inlining file contents into prompts and demonstrates passing a source directory, but provides no instructions to exclude secrets, minimize context, or check whether external disclosure is permitted. Delegation to OpenAI is the skill's stated purpose, which reduces concern about covert exfiltration, but does not ensure that every included file is appropriate to transmit.

Content

No source excerpt is available for this finding.