Back to skill

Security audit

Cluster Agent Swarm

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed platform-operations swarm, but it asks users to install mutable remote content for high-impact cluster and cloud workflows without pinning or integrity checks.

Review this before installing in any environment with real cluster, cloud, GitOps, registry, or repository credentials. Prefer a pinned commit or signed release, verify provenance, install in a least-privileged sandbox first, and only grant production access after confirming the complete referenced skill files and scripts match what was reviewed.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:46
Finding
Unpinned Remote Skill Installation Creates a Supply-Chain Risk## Vulnerability Details **File Location**: `SKILL.md:46-73` **Vulnerability Type**: Unpinned third-party CLI and mutable remote Skill sources **Risk Level**: Medium The installation instructions execute an unversioned npm CLI through `npx` and install Skill content from mutable GitHub repository paths: ```bash npx skills add https://github.com/kcns008/cluster-agent-swarm-skills # Orchestrator - Task routing and coordination npx skills add https://github.com/kcns008/cluster-agent-swarm-skills/skills/orchestrator # Cluster Ops - Atlas (cluster operations) npx skills add https://github.com/kcns008/cluster-agent-swarm-skills/skills/cluster-ops # GitOps - Flow (ArgoCD, Helm, Kustomize) npx skills add https://github.com/kcns008/cluster-agent-swarm-skills/skills/gitops # Security - Shield (RBAC, policies, CVEs) npx skills add https://github.com/kcns008/cluster-agent-swarm-skills/skills/security # Observability - Pulse (metrics, alerts, incidents) npx skills add https://github.com/kcns008/cluster-agent-swarm-skills/skills/observability # Artifacts - Cache (registries, SBOM, promotions) npx skills add https://github.com/kcns008/cluster-agent-swarm-skills/skills/artifacts # Developer Experience - Desk (namespaces, onboarding) npx skills add https://github.com/kcns008/cluster-agent-swarm-skills/skills/developer-experience ``` ### Technical Analysis The command does not pin the `skills` npm package to a reviewed version. It also references GitHub content without an immutable commit hash, verified release artifact, checksum, or cryptographic signature. Consequently, the effective installation behavior and installed Skill instructions may change after this artifact has been audited. The project artifact contains only `SKILL.md`, although its documentation advertises additional agent definitions, scripts, and references. Those remote components were therefore unavailable for review. This does not establish that the u ...[truncated 1897 chars]
Remediation
## Remediation Suggestions 1. Pin the npm CLI to a specifically reviewed version, for example by using an exact package version rather than allowing `npx` to resolve the latest release. 2. Reference the remote repository by an immutable full commit SHA instead of a mutable default branch or directory URL. 3. Vendor the required Skill files, scripts, and references into the audited project so their complete behavior can be reviewed before installation. 4. Publish signed release artifacts and verify their cryptographic signatures or SHA-256 checksums before installation. 5. Use a lockfile and an approved internal package registry or mirror for installation tooling where practical. 6. Run installation in a sandboxed, least-privileged environment without production cluster, cloud, registry, or Git credentials. 7. Add automated dependency and repository integrity monitoring, including alerts for unexpected changes to pinned artifacts. 8. Require security review and provenance verification before updating either the pinned CLI version or repository commit.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (15)

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding
The skill instructs users to run `npx skills add` against a remote GitHub source without any version pinning, commit pinning, or integrity verification. In a platform-operations skill with broad Kubernetes/OpenShift tooling, this creates a supply-chain risk where a later upstream change or repo compromise could deliver altered agent behavior or malicious content into a privileged operational environment.

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding
This line again recommends `npx skills add` from a mutable remote repository path with no pinned version. Because the skill is intended for cluster administration workflows, a compromised or changed upstream source could alter prompts, scripts, or instructions in ways that influence privileged operations.

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding
The installation command for the individual skill is unpinned and fetched from a remote GitHub path. This exposes consumers to repository drift and supply-chain compromise, especially dangerous here because the skill advertises access to infrastructure tools like `kubectl`, `oc`, cloud CLIs, and GitOps tooling.

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding
This unpinned `npx skills add` reference allows installation of whatever content exists at the remote path at execution time. In the context of cluster operations, that could silently change operational guidance or embedded automation affecting upgrades, scaling, or other privileged workflows.

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding
The command installs the GitOps-related skill from a mutable remote location with no version or integrity pin. Since GitOps agents can influence deployment state and reconciliation behavior, a malicious update could steer deployments, introduce unsafe manifests, or manipulate rollback/sync instructions.

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding
This line installs the security-focused skill from an unpinned remote path. A malicious or altered upstream version could subvert audits, suppress warnings, or provide dangerous remediation steps while appearing authoritative in a sensitive security context.

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding
The observability skill is fetched via `npx skills add` without pinning, making the installed content dependent on the current state of a remote repository. In production operations, compromised observability guidance could misdirect incident response, hide meaningful signals, or encourage risky diagnostic actions.

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding
This installation command for another sub-skill is likewise unpinned and mutable. Because the swarm spans multiple operational domains, each unpinned install increases the attack surface for prompt-level or workflow-level supply-chain compromise across cluster, artifact, or developer-experience operations.

Skill Enumeration

Medium
Category
Agent Snooping
Content
## Reference Documentation

For detailed capabilities of each agent, refer to individual SKILL.md files:
- `skills/orchestrator/SKILL.md` - Full Orchestrator documentation
- `skills/cluster-ops/SKILL.md` - Full Cluster Ops documentation
- `skills/gitops/SKILL.md` - Full GitOps documentation
- `skills/security/SKILL.md` - Full Security documentation
Confidence
80% confidence
Finding
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Skill Enumeration

Medium
Category
Agent Snooping
Content
For detailed capabilities of each agent, refer to individual SKILL.md files:
- `skills/orchestrator/SKILL.md` - Full Orchestrator documentation
- `skills/cluster-ops/SKILL.md` - Full Cluster Ops documentation
- `skills/gitops/SKILL.md` - Full GitOps documentation
- `skills/security/SKILL.md` - Full Security documentation
- `skills/observability/SKILL.md` - Full Observability documentation
Confidence
80% confidence
Finding
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Skill Enumeration

Medium
Category
Agent Snooping
Content
For detailed capabilities of each agent, refer to individual SKILL.md files:
- `skills/orchestrator/SKILL.md` - Full Orchestrator documentation
- `skills/cluster-ops/SKILL.md` - Full Cluster Ops documentation
- `skills/gitops/SKILL.md` - Full GitOps documentation
- `skills/security/SKILL.md` - Full Security documentation
- `skills/observability/SKILL.md` - Full Observability documentation
- `skills/artifacts/SKILL.md` - Full Artifacts documentation
Confidence
80% confidence
Finding
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Skill Enumeration

Medium
Category
Agent Snooping
Content
- `skills/orchestrator/SKILL.md` - Full Orchestrator documentation
- `skills/cluster-ops/SKILL.md` - Full Cluster Ops documentation
- `skills/gitops/SKILL.md` - Full GitOps documentation
- `skills/security/SKILL.md` - Full Security documentation
- `skills/observability/SKILL.md` - Full Observability documentation
- `skills/artifacts/SKILL.md` - Full Artifacts documentation
- `skills/developer-experience/SKILL.md` - Full Developer Experience documentation
Confidence
80% confidence
Finding
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Skill Enumeration

Medium
Category
Agent Snooping
Content
- `skills/cluster-ops/SKILL.md` - Full Cluster Ops documentation
- `skills/gitops/SKILL.md` - Full GitOps documentation
- `skills/security/SKILL.md` - Full Security documentation
- `skills/observability/SKILL.md` - Full Observability documentation
- `skills/artifacts/SKILL.md` - Full Artifacts documentation
- `skills/developer-experience/SKILL.md` - Full Developer Experience documentation
Confidence
80% confidence
Finding
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Skill Enumeration

Medium
Category
Agent Snooping
Content
- `skills/gitops/SKILL.md` - Full GitOps documentation
- `skills/security/SKILL.md` - Full Security documentation
- `skills/observability/SKILL.md` - Full Observability documentation
- `skills/artifacts/SKILL.md` - Full Artifacts documentation
- `skills/developer-experience/SKILL.md` - Full Developer Experience documentation
Confidence
80% confidence
Finding
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Skill Enumeration

Medium
Category
Agent Snooping
Content
- `skills/security/SKILL.md` - Full Security documentation
- `skills/observability/SKILL.md` - Full Observability documentation
- `skills/artifacts/SKILL.md` - Full Artifacts documentation
- `skills/developer-experience/SKILL.md` - Full Developer Experience documentation
Confidence
80% confidence
Finding
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Static analysis

No suspicious patterns detected.