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
python izations=["records", "documentation", "knowledge"], preferred_model="gemini-flash", cost_tier=1, max_concurrent=3 ), } # --------------------------------------------------------------------------- # Task Model # --------------------------------------------------------------------------- @dataclass class Task: """A unit of work to be dispatched.""" id: str title: str description: str required_skills: list[str] # e.g. ["security", "audit"] priority: int = 2 # 1=critical, 2=normal, 3=low depends_on: list[str] = field(default_factory=list) # task IDs assigned_to: str = "" status: str = "pending" # pending | dispatched | running | done | failed result: str = "" created_at: float = 0.0 started_at: float = 0.0 completed_at: float = 0.0 cost_estimate: float = 0.0 def to_dict(self) -> dict: return asdict(self) @dataclass class DispatchPlan: """Execution plan for a set of tas