Security audit

OpenClaw Diary Core

Security checks for vulnerabilities and agentic risk

Overview

This diary skill is not clearly malicious, but it can broadly capture personal conversation content, persist it locally, and sync it to Feishu when configured.

Install only if you want an assistant that may save personal thoughts, article discussions, and collaboration notes. Keep Feishu sync disabled unless you are comfortable sending diary content to Feishu, review `personality.file` and storage paths before use, and disable `user_identity` if you do not want the skill reading profile/preference files.

A.I.G

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

Findings (1)

T01 · Skill Instruction Hijacking

Warning

Location: SKILL.md:41
Finding: Unvalidated Personality Path Enables Instruction Hijacking and Arbitrary File Loading
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection

Findings (15)

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 99% confidence
Finding: The description promotes automatic recording and multi-platform sync but does not prominently warn users that their conversation content may be written to local files and sent to external services like Feishu. Lack of upfront disclosure undermines informed consent and can lead to sensitive data being persisted or transmitted unexpectedly.
Content: No source excerpt is available for this finding.

Vague Triggers

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The trigger conditions are so broad that ordinary conversation, time references, or article sharing can cause the skill to activate and potentially store user content without a clear intentional request. In a journaling skill that writes locally and may sync externally, ambiguous activation materially increases the chance of unintentional data capture.
Content: No source excerpt is available for this finding.

Vague Triggers

High

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The boundary between normal discussion and journaling is unclear because sharing thoughts or articles is treated as sufficient to trigger diary behavior. This ambiguity is especially dangerous in a persistence-capable skill, since users may believe they are chatting normally while the agent interprets the exchange as something to store permanently.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The README promotes local and multi-platform syncing for diary content but does not clearly warn users that personal journal data may be transmitted to external services such as Feishu, Flomo, or Notion. Because diary entries often contain highly sensitive personal, emotional, or professional information, insufficient disclosure can cause users to unknowingly expose private data to third parties and broader retention policies.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The README presents a conflicting privacy model: it claims the system only records what the user explicitly says, while also encouraging proactive prompting from vague conversational cues like '今天' and '刚才'. In a diary skill handling sensitive personal content, this ambiguity can lead to unintentional collection and retention of private information because users may not realize casual remarks can trigger recording workflows.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The automatic activation examples are broad enough to match ordinary conversation, such as mentioning 'today', 'just now', emotions, or sharing links. In a diary assistant, overbroad triggering increases the chance of collecting sensitive personal data outside the user's expectations, creating privacy and consent risks even without malicious intent.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The FAQ minimizes collection by saying the AI does not record all conversations and only records limited categories, but earlier sections describe broader automatic triggering from casual conversation and article sharing. This inconsistency can mislead users about when data is captured, which is especially risky for a journaling skill that may store intimate thoughts locally or sync them to third-party platforms.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The skill description and operational instructions are written to enforce Chinese-language behavior, including trigger phrases and user-facing prompts, with only a limited English alias like "journal". There is no opt-in language choice or documented reason that the skill must operate only in Chinese, which can violate language or locale policy requirements.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The skill instructs invoking a shell command (echo $HOME) solely to expand ~, which introduces unnecessary command execution into a workflow that otherwise only needs file-path handling. Even a simple shell dependency expands the agent's operational scope and creates avoidable risk if shell use becomes normalized or is later extended with user-influenced input.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 86% confidence
Finding: The skill reads identity.md and preferences.md files to adjust interaction style, but those files may contain unrelated personal data beyond what is needed to record a diary entry. This broadens collection of sensitive information and creates privacy risk through unnecessary access to user-profile material outside the core journaling action.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 88% confidence
Finding: The script’s user-facing messages and usage instruction are entirely in Chinese, including the suggested invocation phrase at L59. Under the policy, forcing a specific language without offering a language choice or documenting a justified locale restriction is a natural-language policy violation.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The implicit triggers are broad, everyday phrases such as '刚才' and '我在想', and the skill description explicitly instructs proactive prompting even during casual chat. In a diary skill that stores personal thoughts locally and may sync them to third-party platforms, this can cause unintended activation, over-collection of sensitive personal data, and surprise persistence of conversational content.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 72% confidence
Finding: Several comments and descriptions are written only in Chinese, while some trigger phrases are in English, but the file does not state any user-selectable language or locale behavior. This can amount to an implicit language policy that is not clearly documented or offered as an opt-in choice.
Content: No source excerpt is available for this finding.

Vague Triggers

Low

Category: Not specified by scanner
Confidence: 76% confidence
Finding: The tags "主动记录", "陪读", and "协作" are high-level concepts rather than precise invocation criteria, which can make the skill's activation boundaries unclear when combined with broad triggers. This increases the risk of unintended invocation because the config does not explain how these tags limit or exclude activation contexts.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 82% confidence
Finding: This markdown file primarily presents the skill in Chinese, but the example interactions at L111-L119 switch to English responses and user prompts. That introduces an implicit language behavior without stating that the user can choose their preferred language, which can conflict with language/locale policy expectations.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.