Back to skill

Security audit

Super Auto Updater Hold

Security checks for vulnerabilities and agentic risk

Overview

This skill is an auto-updater, but it schedules unattended changes to Clawdbot and every installed skill, so users should review it carefully before enabling it.

Install only if you intentionally want unattended updates to Clawdbot and every installed skill. Prefer a dry-run or notify-only schedule, require approval before applying updates, pin or review versions where possible, run it under a low-privilege account, keep backups, and adjust the timezone before enabling the cron job.

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • System PersistenceInstalls backdoors, hooks, services, or scheduled tasks that survive the run
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (2)

T06 · System Persistence

Error
Location
SKILL.md:27
Finding

Persistent Unattended Auto-Update Task

Content
View full analysis
Remediation
View remediation

T03 · Remote Payload Retrieval and Execution

Error
Location
SKILL.md:51
Finding

Automatic Retrieval and Installation of Mutable Remote Code

Content
View full analysis
/dev/null && npm list -g clawdbot &> /dev/null; then npm update -g clawdbot@latest 2>&1 | tee -a "$LOG_FILE" elif command -v pnpm &> /dev/null && pnpm list -g clawdbot &> /dev/null; then pnpm update -g clawdbot@latest 2>&1 | tee -a "$LOG_FILE" elif command -v bun &> /dev/null; then bun update -g clawdbot@latest 2>&1 | tee -a "$LOG_FILE" else log "Running clawdbot update (source install)" clawdbot update 2>&1 | tee -a "$LOG_FILE" || true fi # Run doctor for migrations log "Running doctor..." clawdbot doctor --yes 2>&1 | tee -a "$LOG_FILE" || true # Capture new version CLAWDBOT_VERSION_AFTER=$(clawdbot --version 2>/dev/null || echo "unknown") # Update skills log "Updating skills via ClawdHub..." SKILL_OUTPUT=$(clawdhub update --all 2>&1) || true ``` ### Technical Analysis The workflow automatically retrieves and installs code from external package, source, and skill registries. The use of the mutable `@latest` tag and `clawdhub update --all` means the effective code installed in the future is not the same code that was available during this audit. No exact version pins, cryptographic checksums, signature-verification policy, publisher allowlist, change review, sandbox validation, or user confirmation step is specified. Running `cla ...[truncated 2079 chars]
Remediation
View remediation
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

Self-Modification

High
Category
Rogue Agent
Confidence
97% confidence
Finding

The script performs clawdhub update --all, which is direct self-modification of the agent's skillset from an external source without review. This is dangerous because it expands the agent's capabilities and behavior over time based on upstream content, creating a supply-chain and persistence risk if a skill is compromised, malicious, or simply incompatible.

Content

Scanner excerpt · references/agent-guide.md (reported line 61)May include surrounding context.

md
# Capture new version
CLAWDBOT_VERSION_AFTER=$(clawdbot --version 2>/dev/null || echo "unknown")

# Update skills
log "Updating skills via ClawdHub..."
SKILL_OUTPUT=$(clawdhub update --all 2>&1) || true
echo "$SKILL_OUTPUT" >> "$LOG_FILE"

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill explicitly sets up unattended daily updates that modify Clawdbot and all installed skills, but it does not present a prominent warning that software will be changed automatically on a schedule and without per-update confirmation. Automatic modification of installed software increases supply-chain and stability risk because newly published packages or skills are applied without user review, potentially introducing malicious or breaking changes.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · references/agent-guide.md (reported line 21)May include surrounding context.

bun pm ls -g 2>/dev/null | grep clawdbot && echo "bun-global"

text

## Step 2: Create the Update Script (Optional)

For complex setups, create a helper script at `~/.clawdbot/scripts/auto-update.sh`:

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The guide instructs the agent to perform unattended package and skill updates, run migration/repair commands, and deliver the results, but it does not require an explicit user warning, approval, rollback plan, or staging check before making system changes. This is dangerous because automatic updates can break the bot, alter behavior, or introduce malicious upstream changes, and the cron-delivered workflow normalizes repeated autonomous modification of the environment.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

The example cron command hard-codes the timezone to America/Los_Angeles, which can cause the updater to run at an unexpected local time if copied verbatim by users outside that locale. In a skill that performs unattended software updates, unexpected execution timing makes the behavior less predictable and can increase operational risk, such as updates occurring during active usage or maintenance windows.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

Line L086 specifies that daily updates will run at 4:00 AM in the America/Los_Angeles timezone. Under the policy, locale constraints should either be user-selectable or clearly justified; this example presents a fixed locale without that context.

Content

No source excerpt is available for this finding.