Back to skill

Security audit

Adult Social Skills

Security checks for vulnerabilities and agentic risk

Overview

This is mostly a coherent adult social-skills guidance prompt, but it requests unnecessary filesystem access and uses an unpinned installer command.

Review this before installing. The social advice content itself is purpose-aligned, but the publisher should remove the filesystem requirement or tightly scope it, and should pin the installer version. If you install anyway, do so in a least-privileged environment without exposing unrelated local files or credentials.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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
Findings (2)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:14
Finding

Unpinned Package Execution During Installation

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, line 14
Vulnerability Type: Unpinned third-party installer dependency
Risk Level: Medium

Vulnerable Code:

yaml
install: "npx clawhub install howtousehumans/adult-social-skills"

Technical Analysis

The installation command invokes clawhub through npx without specifying an audited package version or integrity hash. If the package is not already available locally, npx can retrieve the currently resolved version from the configured npm registry and execute it with the installing user's privileges.

Consequently, the code executed during installation can differ from the code that existed when this Skill was reviewed. This creates a supply-chain trust boundary in which compromise of the package publisher, registry account, package resolution process, or an upstream dependency could introduce arbitrary behavior.

Attack Path

  1. An attacker compromises the publisher account, package distribution channel, or another component involved in resolving the unpinned clawhub package.
  2. The attacker publishes a malicious or modified version that remains compatible with the unversioned command.
  3. A user runs the documented npx clawhub install howtousehumans/adult-social-skills command.
  4. npx resolves, downloads, and executes the attacker-controlled package.
  5. The malicious installer performs actions with the privileges and environmental access of the user running the command.

Impact Assessment

Successful exploitation could result in arbitrary code execution under the installing user's account. Depending on that account's privileges and environment, the payload could access local files, environment variables, credentials, project data, and network resources. If installation is performed by a privileged user or in a CI/CD environment, the impact could extend to system-wide modification, deployment credentials, build artifacts, and other con ...[truncated 22 chars]

Remediation
View remediation

Remediation Suggestions

  • Pin clawhub to a specific, reviewed version rather than allowing npx to resolve the latest available release.
  • Use a lockfile or another reproducible dependency-resolution mechanism.
  • Verify package provenance and integrity through registry signatures or cryptographic hashes where supported.
  • Run installation with the least-privileged account necessary and avoid exposing unrelated credentials or sensitive environment variables.
  • Review the pinned package and its transitive dependencies before updating the approved version.
  • In automated environments, install dependencies from an approved internal registry or immutable artifact repository.

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
SKILL.md:11
Finding

Unnecessary Filesystem Capability Violates Least Privilege

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 11-14
Vulnerability Type: Excessive tool permission
Risk Level: Medium

Vulnerable Code:

yaml
openclaw:
  requires:
    tools: [filesystem]
  install: "npx clawhub install howtousehumans/adult-social-skills"

Technical Analysis

The Skill requests access to the general filesystem tool, but its documented behavior consists of conversational guidance about social skills. None of the reviewed instructions requires reading, creating, modifying, or deleting local files.

Granting a broad capability that is unrelated to the Skill's legitimate function violates the principle of least privilege. Although the reviewed Skill does not directly instruct the agent to misuse this tool, making the capability available increases the consequences of prompt injection, unsafe tool composition, agent malfunction, or future unauthorized changes to the Skill.

Attack Path

  1. The Skill is loaded and receives the declared filesystem capability.
  2. The same agent session processes an attacker-controlled or otherwise untrusted instruction, or the Skill is modified in a later release.
  3. The hostile instruction induces the agent to invoke the filesystem tool under the authority granted to this Skill.
  4. The agent reads or modifies files that are unrelated to the social-advice task.
  5. Retrieved information may be exposed through the conversation or local data may be corrupted, replaced, or deleted, subject to the filesystem tool's actual sandbox boundaries.

Impact Assessment

The obtainable privileges depend on the filesystem tool's configured scope and the operating-system identity of the agent. Potential impact includes unauthorized disclosure of readable local files and unauthorized modification or deletion of writable data. If sensitive configuration files, credentials, source code, or user documents are within scope, exploitation could compromise co ...[truncated 81 chars]

Remediation
View remediation

Remediation Suggestions

  • Remove filesystem from the declared tool requirements because no reviewed functionality depends on it.
  • If persistence of the documented session state is necessary, use a narrowly scoped, Skill-specific state store rather than unrestricted filesystem access.
  • Restrict any unavoidable file capability to a dedicated directory with explicit read/write operations and no access to credentials, user documents, or system configuration.
  • Require user confirmation before any operation that reads sensitive data or changes files.
  • Enforce filesystem sandboxing at runtime rather than relying solely on Skill instructions.
  • Add permission reviews to future releases so newly requested tools must be justified by specific documented functionality.
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)

Vague Triggers

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The automation trigger fires when a user 'expresses loneliness, isolation, or having no friends,' which is highly ambiguous and likely to match common emotional disclosures outside the user’s intent to invoke this skill. In assistant systems, overly broad emotional-state triggers can cause unsolicited intervention, context hijacking, and inappropriate routing of sensitive conversations, especially when users may actually need a different support flow or no automation at all.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The description says to use the skill when someone is lonely, has moved to a new city, wants to make friends, struggles in group settings, or needs to rebuild a social life. These conditions are broad, overlap with common conversation topics, and do not include exclusions or scope constraints, which increases the chance of unintended invocation.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The usage guidance includes general situations like having no friends, feeling lonely, or wanting to host gatherings, but it does not define when the skill should not be used. Without narrower constraints or negative examples, these broad descriptions can collide with many ordinary conversations about mood or social life.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The trigger activates when the user asks about navigating parties, work events, or group settings, which are common and varied topics. Because it lacks contextual constraints or exclusion criteria, it could invoke this skill for broad workplace, etiquette, or anxiety-related questions that may need different handling.

Content

No source excerpt is available for this finding.