Back to skill

Security audit

Komandr

Security checks for vulnerabilities and agentic risk

Overview

The skill openly connects to a remote task service, but it gives remote task content broad authority and sends credentials and results to configurable endpoints without enough safeguards.

Install only if you trust the Komandr server and everyone who can create tasks for your agent. Use a tightly sandboxed agent account, avoid sensitive repositories or secrets, pin and review the runtime dependency, and configure only trusted HTTPS Komandr endpoints with revocable API keys.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (3)

T01 · Skill Instruction Hijacking

Error
Location
SKILL.md:51
Finding

Untrusted Remote Task Content Is Treated as Authoritative Agent Instructions

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
scripts/komandr-bridge.ts:22
Finding

Bearer API Key Can Be Transmitted to an Arbitrary or Plaintext Endpoint

Content
View full analysis
{ const url = `${BASE_URL}${path}`; const headers: Record = { Authorization: `Bearer ${API_KEY}`, "Content-Type": "application/json", }; ``` ### Technical Analysis `KOMANDR_URL` is accepted directly from the process environment and used as the destination for every authenticated request. The implementation does not: - Parse and validate the URL. - Require HTTPS. - Restrict destinations to approved hostnames. - Reject embedded URL credentials or abnormal schemes. - Bind a credential to the server for which it was issued. - Ask for confirmation before sending the credential to a custom server. Every bridge command attaches `KOMANDR_API_KEY` as a bearer token. If `KOMANDR_URL` points to an attacker-controlled HTTPS endpoint, that endpoint directly receives the credential. If it uses plaintext HTTP, the credential can also be exposed to network observers or intermediaries. Sending an authentication token to the intended Komandr server is necessary for the declared functionality. Allowing that token to be sent to any destination without validation exceeds the minimum safe privilege needed to implement that functionality. ### Attack Path 1. An attacker modifies the agent’s environment configuration or persuades a user to set `KOMANDR_URL` to an attacker-c ...[truncated 1264 chars]
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
scripts/komandr-bridge.ts:1
Finding

Unpinned Runtime Package Execution Through npx

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (40)

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 30)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 33)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 36)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 39)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 42)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 45)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 48)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 60)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 68)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 80)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 90)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 91)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 92)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 102)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 112)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 122)May include surrounding context.

md
npx tsx scripts/komandr-bridge.ts me

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
98% confidence
Finding

The skill explicitly instructs the agent to use environment variables and make network requests, but it does not declare any tool scope such as permissions or allowed-tools. This weakens least-privilege controls and can cause the host to grant broader access than users expect for a skill that polls remote tasks and submits results externally.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
84% confidence
Finding

The skill establishes an ongoing agent-to-service relationship with polling, acceptance, progress updates, submissions, and periodic heartbeats, effectively persisting operational state with a remote service. In context this is intentional, but it increases risk because a remote party can continuously feed tasks and receive outputs over time, expanding the blast radius of any misuse or compromise.

Content

Scanner excerpt · SKILL.md (reported line 8)May include surrounding context.

md
# Komandr Skill

This skill connects you to the **Komandr Command Center**, a task orchestration platform. Humans create tasks in Komandr and assign them to AI agents. Your job is to poll for tasks, accept them, do the work, and submit results.

## Installation

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill causes task descriptions, context, progress messages, summaries, results, and optional artifacts to be sent to an external Komandr server, but the description does not prominently warn users about that data flow. This can lead to unintended disclosure of sensitive repository data, prompts, outputs, or operational metadata to a third party.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

Using npx tsx without pinning an exact package version allows the runtime to fetch whatever version is current at execution time. If the package or dependency chain is compromised or a breaking version is published, the agent may execute untrusted code while handling API keys and remote task data.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

Using npx tsx without pinning an exact package version allows the runtime to fetch whatever version is current at execution time. If the package or dependency chain is compromised or a breaking version is published, the agent may execute untrusted code while handling API keys and remote task data.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

Using npx tsx without pinning an exact package version allows the runtime to fetch whatever version is current at execution time. If the package or dependency chain is compromised or a breaking version is published, the agent may execute untrusted code while handling API keys and remote task data.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

Using npx tsx without pinning an exact package version allows the runtime to fetch whatever version is current at execution time. If the package or dependency chain is compromised or a breaking version is published, the agent may execute untrusted code while handling API keys and remote task data.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

Using npx tsx without pinning an exact package version allows the runtime to fetch whatever version is current at execution time. If the package or dependency chain is compromised or a breaking version is published, the agent may execute untrusted code while handling API keys and remote task data.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

Using npx tsx without pinning an exact package version allows the runtime to fetch whatever version is current at execution time. If the package or dependency chain is compromised or a breaking version is published, the agent may execute untrusted code while handling API keys and remote task data.

Content

No source excerpt is available for this finding.