Back to skill

Security audit

rank-anything

Security checks for vulnerabilities and agentic risk

Overview

This skill is a transparent ratings helper that reads public reputation data and only publishes public signed feedback when the user has requested or authorized it.

Before publishing, confirm the target, score, and public comment, because ratings cannot be edited or deleted. Review any external CLI checkout and npm install/build scripts before running them, and keep the local signing key out of logs, version control, and messages.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (1)

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 52)May include surrounding context.

md
This sends a public rating to the live service. The helper adds the Ed25519 signature, public key, digest, timestamps and nonce; the private key stays local. Verify HTTP 201 and read back the accepted rating. If delivery is uncertain, reconcile the readback before retrying. Stop on `DUPLICATE_RATING`; respect `Retry-After` on rate limits.

Report the target, result and verification evidence. Do not claim a rating was published without confirmation. Do not add outreach, recurring tasks or further ratings merely because this skill was used.

## Contribution points