Back to skill

Security audit

OpenClaw Continuity Pack

Security checks for vulnerabilities and agentic risk

Overview

This skill is a real OpenClaw continuity installer, but it also installs broad persistent agent rules, automatic memory behavior, proactive service checks, and high-privilege example settings that deserve manual review.

Install only after reviewing the workspace files it will add. Prefer the continuity-only route, avoid --force unless you have backups and a reviewed diff, do not copy the example config with exec ask off/full security into a sensitive environment, and treat the missing runtime patch as a blocker for the full continuity route.

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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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 (5)

T01 · Skill Instruction Hijacking

Error
Location
assets/workspace/AGENTS.md:7
Finding

Persistent Agent Autonomy and Approval-Policy Hijacking

Content
View full analysis
Remediation
View remediation

T02 · Agent Memory Poisoning

Error
Location
assets/workspace/AGENTS.md:56
Finding

Autonomous Poisoning and Deletion of Persistent Agent Memory

Content
View full analysis
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Error
Location
assets/config/openclaw.example.json:47
Finding

Example Configuration Enables Unrestricted Gateway Execution Without Approval

Content
View full analysis
": [ "" ] } }, "exec": { "host": "gateway", "security": "full", "ask": "off" } } ``` ### Technical Analysis The configuration example combines four high-risk settings: - A full tool profile. - Elevated functionality enabled. - Command execution on the gateway host. - Full execution security with approval prompts disabled. Continuity state management does not require unrestricted gateway command execution or elevated tools. This configuration therefore violates least privilege and materially increases the effect of prompt injection, memory poisoning, configuration mistakes, or compromised workspace instructions. The installer does not automatically copy this configuration into the user's live configuration, which reduces immediate exploitability. However, it is supplied and documented as a sanitized continuity-oriented example, making it likely that users will adopt it as a recommended baseline. ### Attack Path 1. A user copies or adapts `openclaw.example.json` as instructed by the package documentation. 2. The user replaces the placeholder channel and sender values but retains the full/elevated execution settings. 3. The agent loads the bundled persistent workspace instructions. 4. A malicious prompt, poisoned memory entry, untrusted repository file, or external message induces a tool call. 5. Because execution is hosted on the gateway with `security: "full"` and `ask: "off"`, the command can run without interactive confirmation. 6. The command operates with the gateway process's filesystem, network, credential, and repository permissions. ### Impact As ...[truncated 592 chars]
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
assets/workspace/AGENTS.md:201
Finding

Continuity Package Directs Unrelated Reconnaissance of Sensitive Services

Content
View full analysis
8h since you said anything ``` ```markdown **Proactive work you can do without asking:** - Read and organize memory files - Check on projects (git status, etc.) - Update documentation - Commit and push your own changes - **Review and update MEMORY.md** (see below) ``` ### Technical Analysis The declared purpose of the package is continuity and session rollover. Polling email, calendars, social-media notifications, weather services, repositories, and private memory is not necessary to maintain task continuity. These directives encourage recurring access to sensitive connected services without task-specific authorization. If the OpenClaw deployment has authenticated email, calendar, social, or repository integrations, the agent may retrieve private content solely because a heartbeat occurred. No code in the audited package directly implements credential theft or data exfiltration. The risk arises from persistent instructions directing an already privileged agent to use connected services beyond the Skill's legitimate functional scope. ### Attack Path 1. The workspace installer copies the bundled `AGENTS.md` into th ...[truncated 1106 chars]
Remediation
View remediation

T01 · Skill Instruction Hijacking

Error
Location
scripts/bootstrap_workspace.py:27
Finding

Force Installation Can Overwrite Governing Instructions and User Context

Content
View full analysis
str: dst.parent.mkdir(parents=True, exist_ok=True) if dst.exists() and not force: return f"SKIP {dst} (exists)" shutil.copy2(src, dst) return f"COPY {src.name} -> {dst}" ``` The full profile includes governing instructions, identity, user context, and infrastructure notes: ```python full_file_map = { assets_root / "SOUL.md": workspace / "SOUL.md", assets_root / "HEARTBEAT.md": workspace / "HEARTBEAT.md", assets_root / "AGENTS.md": workspace / "AGENTS.md", assets_root / "SESSION_CONTINUITY.md": workspace / "SESSION_CONTINUITY.md", assets_root / "plans" / "TEMPLATE.md": workspace / "plans" / "TEMPLATE.md", assets_root / "status" / "TEMPLATE.md": workspace / "status" / "TEMPLATE.md", assets_root / "handoff" / "TEMPLATE.md": workspace / "handoff" / "TEMPLATE.md", assets_root / "memory" / "README.md": workspace / "memory" / "README.md", assets_root / "temp" / "README.md": workspace / "temp" / "README.md", assets_root / "USER.template.md": workspace / "USER.md", assets_root / "TOOLS.template.md": workspace / "TOOLS.md", } ``` The command-line interface exposes the overwrite operation as one global flag: ```python parser.add_argument("--force", action="store_true", help="Overwrite existing files") ``` ### Technical Analysis The installer does not overwrite existing files by default, which is a useful safeguard. However, a single `--force` flag disables that safeguard for every mapped file. There is no backup, content diff, per-file confirmation, atomic rollback, or restriction to continuity-specific records. ...[truncated 1703 chars]
Remediation
View remediation
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
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (34)

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

A second description-behavior mismatch finding indicates the skill may function only as validation or anchor checking rather than as an installer or patching utility as advertised. Misrepresenting a skill that touches source trees and deployment workflows can lead users to run follow-on commands or manual steps under false assumptions, increasing the risk of unsafe edits, broken deployments, or accidental policy bypass.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

A second description-behavior mismatch finding indicates the skill may function only as validation or anchor checking rather than as an installer or patching utility as advertised. Misrepresenting a skill that touches source trees and deployment workflows can lead users to run follow-on commands or manual steps under false assumptions, increasing the risk of unsafe edits, broken deployments, or accidental policy bypass.

Content

No source excerpt is available for this finding.

Self-Modification

High
Category
Rogue Agent
Confidence
85% confidence
Finding

Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Content

Scanner excerpt · assets/workspace/TOOLS.template.md (reported line 36)May include surrounding context.

md
## Why Separate?

Skills are shared. Your setup is yours. Keeping them apart means you can update skills without losing your notes, and share skills without leaking your infrastructure.

Add whatever helps you do your job. This is your cheat sheet.

Self-Modification

High
Category
Rogue Agent
Confidence
85% confidence
Finding

Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Content

Scanner excerpt · scripts/bootstrap_workspace.py (reported line 126)May include surrounding context.

python
default="auto",
        help="Target OpenClaw workspace path. Default: auto-detect from env/current workspace, else ~/.openclaw/workspace",
    )
    parser.add_argument("--force", action="store_true", help="Overwrite existing files")
    parser.add_argument(
        "--profile",
        choices=("full", "continuity"),

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill advertises installation and patching workflows that imply shell, file read/write, and environment access, but the manifest does not declare any explicit tool scope or permissions boundary. In an agent ecosystem, missing scope declarations can cause operators and automated policy layers to underestimate what the skill may do, increasing the chance of unintended source modification or command execution.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill references a full continuity install path that applies a runtime patch, rebuilds, and redeploys against a matching source tree, but it does not provide a prominent upfront warning about these source modification and deployment-impacting actions. Users may invoke destructive or environment-altering steps without understanding that application code will be changed and rebuilt, which is especially risky in production-like environments.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The documentation states that the runtime patch enables silent continuity preparation and suppresses visible continuity/context hint messages, which changes user-visible chat behavior, but it does not frame this as a clear behavioral-risk warning. Hidden or less visible system behavior can affect user expectations, transparency, debugging, and auditability, particularly where conversation continuity is being preserved without explicit prompts.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The manifest text uses very broad activation language such as '全自动新对话续接' and '一键安装' without clearly constraining when the skill should be invoked or what user consent is required. Combined with allow_implicit_invocation: true, this increases the chance the agent will trigger the skill in unrelated contexts and perform installation-oriented actions or continuity patch guidance the user did not explicitly request.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The visible metadata and default behavior are written only in Chinese and do not indicate any user language negotiation or documented locale restriction. This can cause users to misunderstand what the skill will do, especially because it discusses installation paths and runtime patch behavior, reducing informed consent and increasing the risk of accidental invocation or misuse.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This configuration enables automatic memory flushing that silently writes durable state to workspace files and instructs the agent to reply with NO_REPLY unless absolutely necessary. That creates a covert persistence channel for conversation-derived data, increasing the risk of storing sensitive information without user awareness or meaningful consent.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The elevated-tool activation scope is defined with generic placeholders for both channel and sender, which provides no concrete security boundary in the example configuration. In a skill that explicitly enables elevated tools, an operator may replace these loosely or incorrectly, causing privileged actions to be granted to unintended message sources and enabling command execution or other sensitive operations.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The exec tool is configured with full security, hosted on the gateway, and prompting turned off, allowing shell or execution actions to proceed without interactive approval. In the context of a continuity/automation skill that aims for silent background continuation, this materially increases the chance of unreviewed command execution, misuse of credentials, and destructive system actions.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
88% confidence
Finding

The instruction not to ask the user for harmless reads or dry-runs can be acceptable in tightly sandboxed local environments, but as a reusable skill it weakens consent boundaries and normalizes silent tool execution. Combined with the document's broader autonomy themes, it may cause an agent to act with insufficient confirmation about environment, scope, or data sensitivity.

Content

Scanner excerpt · assets/workspace/AGENTS.md (reported line 24)May include surrounding context.

md
- what remains
   - the only blocker, if any
7. Never create or modify AGENTS.md again just because the user repeats this preference. Treat this rule as persistent.
8. If this local Control UI session already has exec configured as ask=off and security=full, do not ask the user to paste /approve for harmless local reads or dry-runs. Execute them directly and only mention /approve when the tool explicitly returns approval-pending.

## First Run

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The activation condition says to treat any task with more than one phase, more than one workstream, or likely to take more than about 10 minutes as a long task. Those criteria are broad and subjective, and the file does not provide negative examples or tighter scope boundaries, which could cause unintended invocation of the long-task workflow.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% 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 · assets/workspace/AGENTS.md (reported line 124)May include surrounding context.

md
## Red Lines

- Don't exfiltrate private data. Ever.
- Don't run destructive commands without asking.
- `trash` > `rm` (recoverable beats gone forever)
- When in doubt, ask.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The heartbeat section authorizes proactive checks of external services and even 'commit and push your own changes' without a strong per-action consent boundary. That expands the agent from reactive assistant to autonomous actor with outbound network behavior, creating risk of unintended disclosure, unauthorized communications, or repository changes performed outside explicit user intent.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
95% confidence
Finding

The instruction explicitly permits proactive work 'without asking,' including checking projects and committing/pushing changes. In skill context, this is more dangerous because the same document also encourages heartbeat-driven background activity and external checks, increasing the chance of silent modifications or outbound actions that the user did not explicitly request in the current session.

Content

Scanner excerpt · assets/workspace/AGENTS.md (reported line 262)May include surrounding context.

md
- Nothing new since last check
- You just checked <30 minutes ago

**Proactive work you can do without asking:**

- Read and organize memory files
- Check on projects (git status, etc.)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The activation criteria are broad and subjective (for example, 'likely to touch multiple files/systems' or 'anything the user may want resumed later'), which can cause the protocol to trigger on ordinary work without clear user intent. In this skill's context, that increases the chance of unnecessary persistence and continuity-file creation, expanding data retention and making subsequent automated behavior less predictable.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The protocol directs the agent to automatically append and create/update memory, plan, and status files immediately on task intake, but provides no requirement to notify the user or obtain consent. In a continuity skill, this is especially risky because it silently persists task details and decisions to disk, which may capture sensitive prompts, operational context, or user data beyond the user's expectations.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The file's core instruction is written as a direct persona/specification in Chinese, telling the assistant what it is, but it does not indicate that Chinese is optional or user-selected. This creates a language/locale policy concern because the skill appears to impose a specific language context without opt-in or justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The file forces a specific language for operational instructions and does not provide an opt-in, alternative language, or justification for the locale restriction. Under the stated policy, fixed language behavior without user choice or documented rationale is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This markdown file uses Chinese throughout and does not provide any indication that the user can choose another language or that the document is intentionally limited to a Chinese-speaking audience. Under the stated policy, forcing a specific language without opt-in is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

This markdown file uses only Chinese for critical installation and deployment guidance, which can amount to a language policy violation when no user opt-in or alternative locale is offered. The document does not indicate that the skill is China-specific or otherwise justified to require Chinese-only instructions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The document begins entirely in Chinese and presents the skill usage instructions in that language, with no indication that other languages are supported or that the language choice is optional. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless clearly justified as region-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

This markdown file contains operational instructions primarily in Chinese, but also includes English-only prompt examples and interface terminology such as 'Please reply with READY only.' and 'Control UI'. That effectively imposes a mixed language/locale expectation without any opt-in or explanation, which can conflict with organizational language-choice policies.

Content

No source excerpt is available for this finding.