Back to skill

Security audit

New Relic CLI Skills

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this New Relic CLI skill suspicious because it flags broad activation and possible exposure of commit details and developer identities; some referenced content was only partly analyzed.

Consider the suspicious verdict before installing. Review the broad activation behavior and the CI example's handling of commit titles and developer identities, and account for the incomplete analysis of referenced content. The scan does not establish malicious intent, but it also does not confirm that all content is safe.

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (11)

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 56)May include surrounding context.

md
override the selected profile. See nrql/SKILL.md for preflight and write gates.

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 77)May include surrounding context.

md
**Need to query data or manage organization-scoped NRQL macros?** → [`nrql/SKILL.md`](nrql/SKILL.md)

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 184)May include surrounding context.

md
| [`nrql/`](nrql/SKILL.md) | Custom queries, dashboards, ad-hoc exploration, and approval-gated organization macros |

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 81)May include surrounding context.

md
**Infrastructure/host issues?** → [`infrastructure/SKILL.md`](infrastructure/SKILL.md)

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 188)May include surrounding context.

md
| [`infrastructure/`](infrastructure/SKILL.md) | Host metrics, CPU/memory, process monitoring |

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: network, shell.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Activation conditions include generic topics without requiring a New Relic context.

Medium
Category
Trigger Abuse
Confidence
90% confidence
Finding

The activation criteria include generic requests about users, permissions, performance, and setup without requiring an established New Relic context. This can route unrelated tasks into credentialed New Relic workflows and cause unnecessary access to organizational data, although the explicit write-approval requirements substantially limit the consequences.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
85% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 24)May include surrounding context.

md
macro creation/update/deletion, plus irreversible AI Monitoring dataset/version deletion.
      Dataset deletion cascades to all versions; last-version deletion also deletes its dataset.
      Require explicit approval of exact dataset/version IDs and cascade scope, fresh
      read-back before deletion, and exact-result verification afterward; do not bypass
      confirmation for automation. See aimonitoring/references/datasets.md. Require explicit approval before macro writes,
      inspect downstream query/dashboard impact with coverage gaps disclosed, and read
      the exact target back. Macros are not supported in alert conditions; cached query

Unbounded Output

Medium
Category
Output Handling
Confidence
60% confidence
Finding

Output size or generation rate is not bounded. Unbounded output enables denial-of-service through resource exhaustion, log flooding, or context-window stuffing.

Content

Scanner excerpt · aimonitoring/references/datasets.md (reported line 106)May include surrounding context.

md
Both commands follow `nextCursor` until empty, with a hard cap of **1,000 pages**.
At the cap they silently return accumulated results even if another cursor exists;
there is no truncation warning or continuation flag in the printed output. A later
page error fails the operation rather than returning the earlier partial results.
Only the expected dataset/version entity types are retained. The fixed queries are
`type = 'AIM_DATASET'` and `type = 'AIM_DATASET_VERSION'` respectively.

CI integration sends commit titles and user identities to New Relic without a privacy warning.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The CI example automatically forwards raw commit titles and developer login names to an external observability service. Although deployment metadata is appropriate for this skill, titles can contain confidential issue details or accidentally embedded secrets, and the example supplies no data-minimization or disclosure gate; shell quoting and control-character filtering do not prevent that exposure.

Content

No source excerpt is available for this finding.

Scope Creep

Low
Category
Excessive Agency
Confidence
70% confidence
Finding

Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Content

Scanner excerpt · nrql/SKILL.md (reported line 9)May include surrounding context.

md
## Organization-Scoped Macros

A macro is a named expression fragment referenced as `$macroName` in NRQL. Macros
are shared across accounts in an organization; they are not limited to the account
selected by `--accountId`. Use a User API key with organization-level macro access:
**Organization product admin** or **Organization manager** for create/update/delete;
**Organization read only** grants reads. Never print or paste the key. Macros can