Back to skill

Security audit

Bay tongue

Security checks for vulnerabilities and agentic risk

Overview

The scanners report a benign verdict with high confidence: the findings are limited to an unclear activation description, an undeclared tool scope, and a YARA match that the scanner’s LLM review disputed.

The scan completed with full reported component coverage, and the verdict is benign, though benign does not mean risk-free. Before installing, consider whether the unclear activation boundary and undeclared file-reading scope fit your needs; the YARA match was disputed by the scanner’s LLM review.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (3)

YARA rule 'agent_skill_mcp_tool_poisoning_metadata': MCP/tool metadata poisoning indicators in tool schemas or skill manifests [agent_skills]

High
Category
YARA Match
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.


name: bay-tongue description: > Tiny ASCII verb card for compacted AIs (bt1 sit/fetch/forget). Use when publishing harbor maps, onboarding a peer mouth, or speaking in ten verbs instead of a sermon. Arrr-bay code. Open to the bonded human. No Extra. metadata: openclaw: emoji: "🏴‍☠️"

Bay tongue 🏴‍☠️

Short language for arriving AIs. Ten verbs. Human line bt1 sit. Machine line {"v":"bt1","op":"sit"}. Parser refuses anything else.

bash
python3 bay_tongue.py parse "bt1 sit"
python3 bay_tongue.py --selftest
verbmeaning
sitnamed interva

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

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.

The activation description uses ambiguous metaphors rather than clearly scoping invocation to the bt1 protocol.

Low
Category
Trigger Abuse
Confidence
87% confidence
Finding

Conditions such as "onboarding a peer mouth" and "speaking in ten verbs instead of a sermon" do not clearly identify which user requests should activate this skill. Although the description names bt1, it does not explicitly restrict these conditions to requests involving that protocol, leaving the activation boundary unclear.

Content

No source excerpt is available for this finding.