Back to skill

Security audit

Kubernetes

Security checks across malware telemetry and agentic risk

Overview

This Kubernetes skill is a disclosed local operations helper that keeps cluster notes on the user's machine and avoids storing credentials.

Install only if you are comfortable with the agent keeping local Kubernetes operational notes under `~/Clawic/data/`, including cluster topology, hostnames, RBAC gaps, incident history, and runbooks. Keep that directory private, review it periodically, and do not paste secrets unless you are relying on the skill's redaction rules.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (10)

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The skill instructs the agent to modify shared memory files outside the narrow scope of Kubernetes ingress debugging, including `domains.md` and `memory.md`. That creates a cross-scope side effect where a troubleshooting skill can alter durable operational records, increasing the risk of unauthorized or incorrect changes to shared state that other skills may trust.

Context-Inappropriate Capability

Medium
Confidence
92% confidence
Finding
Embedding registrar and certificate inventory maintenance into an ingress troubleshooting document expands the skill from diagnostics into persistent operational bookkeeping. Because these shared records may be consumed by other workflows, inaccurate or unnecessary updates can propagate bad state, mislead responders, or overwrite authoritative ownership information.

Context-Inappropriate Capability

Medium
Confidence
93% confidence
Finding
The file instructs the agent to persist troubleshooting outputs and cluster details under a fixed local path in the user's home directory, which extends beyond ephemeral Kubernetes debugging into cross-session data retention. That can cause unintended storage of sensitive operational metadata such as network policies, enforcement status, and pod/service CIDRs without user consent or clear scoping.

Description-Behavior Mismatch

Medium
Confidence
91% confidence
Finding
The networking guidance shifts into maintaining a durable knowledge base (`artifacts` and `memory.md`) about future sessions and cluster inventory, which is outside the immediate task of debugging or manifest review. This creates a channel for accumulating environment intelligence over time, increasing exposure if the host is shared, compromised, or the stored data is later reused inappropriately.

Context-Inappropriate Capability

Medium
Confidence
94% confidence
Finding
The document instructs the agent to persist RBAC analysis artifacts to local files and update shared memory, which goes beyond providing RBAC guidance and creates side effects on the host environment. Because these artifacts may contain service account names, permissions, gaps, and operational context, they can expose sensitive security information and create cross-task data leakage or unauthorized state modification.

Context-Inappropriate Capability

Medium
Confidence
88% confidence
Finding
The document instructs the agent to persist operational Kubernetes data into local files under paths like `~/Clawic/data/k8s/memory.md`, `deploys/<year>.md`, and `artifacts/decision-<kebab>.md`, which goes beyond transient debugging and manifest review. In an agent setting, this can cause unauthorized modification of local workspace content, accidental persistence of sensitive operational details, and scope creep from analysis into state-changing file writes.

Missing User Warnings

Low
Confidence
88% confidence
Finding
The markdown tells the user/agent to save data to a specific filesystem path but does not warn that this writes potentially sensitive user and cluster information to disk. Even if the content is operationally useful, silent persistence can violate user expectations and leave residual data behind.

Missing User Warnings

Low
Confidence
90% confidence
Finding
The markdown tells the agent to save RBAC artifacts and update memory without notifying the user that local data will be modified. Hidden persistence is risky even when the content is operationally useful, because it violates least surprise and can cause unauthorized writes or retention of security-relevant information.

Ssd 3

Medium
Confidence
93% confidence
Finding
Persisting detailed RBAC artifacts and known gaps in shared memory can retain sensitive operational context that may later be surfaced to unrelated tasks or users. In a Kubernetes security skill, this is more dangerous because the saved data can map privilege boundaries, exceptions, and service account capabilities that materially aid reconnaissance or privilege escalation.

Session Persistence

Medium
Category
Rogue Agent
Content
- ~/clawic/k8s/
---

**Data.** At the start of every session, read `~/Clawic/data/k8s/config.yaml` (what the user declared) and `~/Clawic/data/k8s/memory.md` (what you observed, plus its `## Boxes` index and `## Due` table). Open any file `## Boxes` names the moment the condition on its line applies — that index is the list of files that exist, never a list you carry in your head. Check `## Due` against today's date and state any overdue item in one line: a statement, not a question. Read `~/Clawic/data/servers/servers.md` before any capacity, sizing, upgrade, or "what do we run" question. If none of it exists, work from defaults and say nothing about it. If data sits at an old location (`~/k8s/` or `~/clawic/k8s/`), move it to `~/Clawic/data/k8s/`, and say in one line that you moved it and from where. Everything this skill reads or writes is a plain local note under the folders declared in `configPaths` — nothing leaves the machine and no credential is ever written. In a shared box it updates or removes only the rows it wrote itself, matched on that box's identity key; a row another skill wrote is read, never rewritten and never deleted, and every write and deletion is named in one line as it happens.

**Write before the session ends** whenever it produced something durable: a cluster discovered, upgraded or retired; a workload sized from real observation; an incident whose cause was finally named; a deploy with its rollback digest; a drill that was timed; an audit finding the user chose to accept; a hostname the cluster now serves; or anything they will want to read again — a runbook, a policy or manifest that finally worked, an architecture decision. `memory-template.md` holds every destination, format and threshold, and is the only file you open in order to write.
Confidence
94% confidence
Finding
write and deletion is named in one line as it happens. **Write before the session ends** whenever it produced something durable: a cluster discovered, upgraded or retired; a workload sized from real

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.