Back to skill

Security audit

CareMax Indicators

Security checks for vulnerabilities and agentic risk

Overview

The skill does expected CareMax health-data work, but it depends on unpinned external auth scripts and can write persistent health readings, so it needs careful review before install.

Install only if you trust the CareMax service and the `KittenYang/caremax-skills` source. Prefer a pinned, reviewed version of `caremax-auth`, and require the agent to confirm metric, value, unit, date, and whose profile before any quick-log write.

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:9
Finding
Unpinned Third-Party Skill Installation Creates a Supply-Chain Risk<![CDATA[ ## Vulnerability Details **File Location**: `SKILL.md`, line 9 **Vulnerability Type**: Unpinned and mutable third-party dependency installation **Risk Level**: Medium ### Vulnerable Code ```markdown > **Requires `caremax-auth` as a sibling directory** (`../caremax-auth/`, same layout as `skills/caremax-auth` + `skills/caremax-indicators` in this repo, or under `~/.agents/skills/`). If missing: `npx skills add KittenYang/caremax-skills` and select caremax-auth. ``` ### Technical Analysis The installation instruction invokes the `skills` npm command through `npx` without pinning the package to a reviewed version. It also identifies the `KittenYang/caremax-skills` repository without pinning it to a specific commit hash or immutable signed release. Both components can therefore resolve to content that differs from what was available during this audit. The installed `caremax-auth` dependency is subsequently trusted to provide shell scripts used for authentication and API access, including `api-call.sh`, `auth-flow.sh`, `list-system-presets.sh`, and `quick-log.sh`. Those scripts are not included in the audited project, so their integrity and behavior cannot be verified here. This is a supply-chain weakness rather than evidence that the current upstream dependency is malicious. Exploitation would require compromise or malicious modification of the npm package, its publishing account, the referenced repository, or another mutable dependency in the installation path. ### Attack Path 1. The required sibling `caremax-auth` directory is absent. 2. An agent or operator follows the documented recovery instruction and executes: ```bash npx skills add KittenYang/caremax-skills ``` 3. `npx` resolves an unpinned version of the `skills` package, and the installer retrieves content from a repository reference that is not bound to an audited commit. 4. If either source has been compromised or maliciously modified, attacker-controlled skill files or script ...[truncated 1233 chars]
Remediation
<![CDATA[ ## Remediation Suggestions 1. Pin the npm CLI dependency to a reviewed exact version rather than allowing `npx` to resolve the latest release: ```bash npx skills@<audited-exact-version> add ... ``` 2. Pin the skill repository to a verified immutable commit hash or a signed release instead of relying on a mutable repository branch or default reference. 3. Publish and verify cryptographic checksums or signatures for the installed `caremax-auth` files before executing any script. 4. Vendor the minimum required, reviewed authentication scripts with the project where licensing and maintenance requirements permit it. 5. Add a verification step that rejects unexpected files or hash changes in `../caremax-auth/scripts/`. 6. Run authentication and API scripts with least privilege, restricting filesystem access, environment variables, credential exposure, and network destinations where practical. 7. Document the exact audited dependency versions and establish a controlled process for reviewing and updating them. ]]>
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (6)

Direct Prompt Extraction

High
Category
System Prompt Leakage
Content
# 3. Present with values and reference ranges
```

## Display Guidelines

- Always show values with units (e.g., "98 μmol/L" not just "98")
- Include reference ranges when available
Confidence
85% confidence
Finding
Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding
Suspicious Unicode normalization or mixed-script content

Vague Triggers

Medium
Confidence
94% confidence
Finding
The description mixes broad intents ('health metrics', 'trend', 'quick log') with trigger terms but does not clearly state when the skill should or should not activate. That ambiguity increases the chance the agent invokes this skill in borderline contexts and unnecessarily accesses or prepares to modify sensitive medical data.

Vague Triggers

Medium
Confidence
86% confidence
Finding
The description mixes broad intents ('health metrics', 'trend', 'quick log') with trigger terms but does not clearly state when the skill should or should not activate. That ambiguity increases the chance the agent invokes this skill in borderline contexts and unnecessarily accesses or prepares to modify sensitive medical data.

Rp1

Medium
Category
MCP Rug Pull
Confidence
89% confidence
Finding
The instruction to run `npx skills add KittenYang/caremax-skills` references an unpinned package/source, which can cause users to fetch whatever current version is published at execution time. This creates a supply-chain risk: a compromised upstream release or repo change could introduce malicious code or altered auth/API scripts into a workflow that handles health data and OAuth tokens.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill advertises 'quickly add a single reading' without a prominent warning that this writes health data to the user's or a family member's persistent profile. Because the data is medical in nature and the skill supports `--member`, insufficient write/transparency controls can lead to accidental record modification, privacy issues, or incorrect entries in another person's profile.

Static analysis

No suspicious patterns detected.