Back to skill

Security audit

VCF SDDC Manager Errors

Security checks for vulnerabilities and agentic risk

Overview

This skill is purpose-aligned for VCF log troubleshooting, but it uses a bearer token with unsafe TLS settings and an overly broad log query that may expose more operational logs than users expect.

Review this skill before installing. Only use it with a least-privilege, read-only Log Insight token, remove the TLS bypass or configure a trusted CA, and narrow the query or activation language so it retrieves only the intended SDDC Manager or vCenter logs after clear user intent.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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)

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:22
Finding

TLS Certificate Verification Disabled for Authenticated API Request

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 22–25
Vulnerability Type: TLS certificate validation bypass
Risk Level: High

bash
curl -s -k -X GET "https://$LOGINSIGHT_HOST/api/v1/events?keyword=error&limit=10" \
  -H "Authorization: Bearer $LOGINSIGHT_API_TOKEN" \
  -H "Accept: application/json" | jq '.events[]? | {timestamp, source, text}'

Technical Analysis

The -k option instructs curl to skip TLS certificate verification. Consequently, curl will accept an untrusted, expired, hostname-mismatched, or attacker-controlled certificate.

This request transmits LOGINSIGHT_API_TOKEN in an Authorization: Bearer header and retrieves potentially sensitive infrastructure logs. Disabling certificate validation removes server-authentication guarantees and makes the connection vulnerable to man-in-the-middle interception. Although the destination is supplied through LOGINSIGHT_HOST rather than being a hard-coded attacker endpoint, any attacker capable of manipulating DNS, routing, proxy settings, or the host configuration could impersonate the Log Insight service.

Attack Path

  1. The skill obtains LOGINSIGHT_HOST and LOGINSIGHT_API_TOKEN from environment variables.
  2. An attacker gains a network interception position or redirects resolution or routing for LOGINSIGHT_HOST.
  3. The attacker presents a certificate that would ordinarily fail TLS verification.
  4. Because curl is invoked with -k, the invalid certificate is accepted.
  5. The request sends the bearer token to the impersonated endpoint.
  6. The attacker captures and potentially reuses the token against the legitimate Log Insight API.
  7. The attacker can also return fabricated JSON events, causing the agent to present false logs or generate misleading troubleshooting conclusions.

Impact Assessment

Successful exploitation may expose the Log Insight API bearer token and sensitive operational log records. The priv ...[truncated 381 chars]

Remediation
View remediation

Remediation Suggestions

Remove the -k option and require standard TLS certificate and hostname validation:

bash
curl -sS -X GET "https://$LOGINSIGHT_HOST/api/v1/events?keyword=error&limit=10" \
  -H "Authorization: Bearer $LOGINSIGHT_API_TOKEN" \
  -H "Accept: application/json" | jq '.events[]? | {timestamp, source, text}'

If the service uses a certificate issued by a private certificate authority, provision the trusted CA certificate and pass it explicitly with --cacert /path/to/trusted-ca.pem. Do not suppress validation globally.

Additionally:

  • Restrict LOGINSIGHT_HOST to an approved hostname or allowlist.
  • Reject host values containing unexpected URL syntax or control characters.
  • Assign the API token only the minimum read permissions required to retrieve events.
  • Rotate the token if the vulnerable command has been used over an untrusted network.
  • Avoid logging or displaying the bearer token.
  • Fail securely when certificate validation does not succeed.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (4)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
60% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SKILL.md (reported line 13)May include surrounding context.

md
- LOGINSIGHT_HOST
        - LOGINSIGHT_API_TOKEN
      bins:
        - curl
        - jq
    primaryEnv: LOGINSIGHT_API_TOKEN
    emoji: ""

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The activation text is broad enough to trigger on generic troubleshooting or recent error requests, causing the skill to retrieve logs in situations where the user may not have intended log access. Because the skill uses stored credentials to query an external logging system, overbroad activation increases the chance of unnecessary sensitive data exposure.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill instructs the agent to use environment-provided API credentials to fetch operational logs from a remote service without an explicit user-facing warning or confirmation. Retrieved logs may contain sensitive infrastructure details, hostnames, stack traces, or security-relevant errors, so silent network retrieval materially increases disclosure risk.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill claims to focus on SDDC Manager and vCenter, but the actual query only searches for the generic keyword "error" across recent events and does not constrain results to those sources. This can expose unrelated operational logs and broaden data access beyond what a user would reasonably expect from the description.

Content

No source excerpt is available for this finding.