Back to skill

Security audit

Openclaw Skill

Security checks for vulnerabilities and agentic risk

Overview

This skill is a mostly coherent AgentIzzy API integration, but it deserves Review because it combines sensitive call/lead data access with webhooks, phone-number provisioning, and billing checkout actions without strong user-confirmation boundaries.

Install only if you trust AgentIzzy with call recordings, transcripts, caller contact details, lead data, and billing/account workflows. Require explicit human approval before registering accounts, configuring webhooks, provisioning phone numbers, or starting checkout, and avoid passing untrusted values into the helper scripts until their JSON and URL construction are hardened.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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 (4)

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/register.sh:15
Finding

Unescaped agent identifier permits JSON request-body manipulation during registration

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/setup.sh:22
Finding

Multiple setup parameters are inserted into JSON without serialization

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Note
Location
scripts/calls.sh:12
Finding

Unencoded call-query inputs can alter the intended API path and query

Content
View full analysis
Remediation
View remediation
&2 exit 2 } [[ "$LIMIT" =~ ^[0-9]+$ ]] && (( LIMIT >= 1 && LIMIT <= 100 )) || { echo '{"error":"Limit must be between 1 and 100"}' >&2 exit 2 } args=( --silent --show-error --fail-with-body --globoff --get "https://api.agentizzy.com/api/v1/agents/${FACILITY_ID}/calls" --data-urlencode "limit=${LIMIT}" -H "Authorization: Bearer ${AGENTIZZY_API_KEY}" ) [ -n "$SINCE" ] && args+=(--data-urlencode "since=${SINCE}") curl "${args[@]}" ``` Also validate `SINCE` against the exact accepted ISO date or timestamp format. If agent identifiers may contain reserved characters, encode the path segment rather than merely applying a restrictive regular expression. ]]>

T09 · Insecure Skill Coding Practices

Note
Location
scripts/leads.sh:12
Finding

Unencoded lead-query inputs can modify the intended request URL

Content
View full analysis
Remediation
View remediation
&2 exit 2 } [[ "$LIMIT" =~ ^[0-9]+$ ]] && (( LIMIT >= 1 && LIMIT <= 100 )) || { echo '{"error":"Limit must be between 1 and 100"}' >&2 exit 2 } curl --silent --show-error --fail-with-body --globoff --get \ "https://api.agentizzy.com/api/v1/agents/${FACILITY_ID}/leads" \ --data-urlencode "limit=${LIMIT}" \ -H "Authorization: Bearer ${AGENTIZZY_API_KEY}" ``` If the API permits characters outside the validated identifier set, use a standards-compliant URL encoder for the individual path segment. Retain server-side ownership checks because client-side validation is not an authorization control. ]]>
Vulnerability Patterns
  • 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 Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (25)

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The skill claims broad phone-agent functionality, but the file mainly documents shell/network access patterns and dependency on an API key rather than implementing the stated call-handling logic directly. This mismatch is risky because operators may approve or invoke the skill under incomplete assumptions about required permissions, external access, and data exposure.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The skill claims broad phone-agent functionality, but the file mainly documents shell/network access patterns and dependency on an API key rather than implementing the stated call-handling logic directly. This mismatch is risky because operators may approve or invoke the skill under incomplete assumptions about required permissions, external access, and data exposure.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
95% confidence
Finding

The skill declares shell-capable behavior and external API usage but does not specify any explicit tool scope or allowed-tools constraints. In agent environments, this increases the chance of unintended shell execution or network access beyond what a user expects, especially because the skill also handles credentials and sensitive call data.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger text is broad enough that the skill could be invoked in many business-automation situations, including ones where users did not specifically intend phone provisioning, transcript access, or lead-data retrieval. Overbroad invocation increases the risk of unnecessary access to sensitive communications data or unintended external actions.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill prominently exposes transcripts, recordings, summaries, sentiment, and lead data, but does not provide a clear user-facing privacy warning at the point of use. Because this data can contain highly sensitive personal and business information, omission of explicit consent and privacy guidance increases the likelihood of inappropriate collection, transmission, or downstream use.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
85% confidence
Finding

The phrase indicating registration is 'instant' with 'no approval needed' encourages autonomous account creation without a human review checkpoint. In an agentic environment, this raises the risk of unattended third-party account creation, credential issuance, and expansion of external dependencies without informed consent.

Content

Scanner excerpt · SKILL.md (reported line 16)May include surrounding context.

Setup

  1. Register for a free API key (instant, no approval needed):
    bash
    curl -s -X POST https://api.agentizzy.com/api/v1/register \
      -H "Content-Type: application/json" \
    

External Transmission

Medium
Category
Data Exfiltration
Confidence
90% confidence
Finding

This step transmits user-supplied registration data to an external service and results in creation of a new API credential. External transmission is expected for this type of integration, but it is still security-relevant because it performs an outbound state-changing action and initiates trust with a third-party platform.

Content

Scanner excerpt · SKILL.md (reported line 18)May include surrounding context.

  1. Register for a free API key (instant, no approval needed):
    bash
    curl -s -X POST https://api.agentizzy.com/api/v1/register \
      -H "Content-Type: application/json" \
      -d '{"name": "My Agent", "agent_id": "my-agent"}'
    

External Transmission

Medium
Category
Data Exfiltration
Confidence
90% confidence
Finding

This step transmits user-supplied registration data to an external service and results in creation of a new API credential. External transmission is expected for this type of integration, but it is still security-relevant because it performs an outbound state-changing action and initiates trust with a third-party platform.

Content

Scanner excerpt · SKILL.md (reported line 18)May include surrounding context.

  1. Register for a free API key (instant, no approval needed):
    bash
    curl -s -X POST https://api.agentizzy.com/api/v1/register \
      -H "Content-Type: application/json" \
      -d '{"name": "My Agent", "agent_id": "my-agent"}'
    

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

Fetching call records sends an authenticated request to a third-party API and can return transcripts and other sensitive caller data. In this skill context, that is security-relevant because the retrieved data may contain personal information, business details, and potentially regulated communications content.

Content

Scanner excerpt · SKILL.md (reported line 158)May include surrounding context.

md
-H "Authorization: Bearer $AGENTIZZY_API_KEY"

# Get calls for an agent
curl -s "https://api.agentizzy.com/api/v1/agents/AGENT_ID/calls?limit=10" \
  -H "Authorization: Bearer $AGENTIZZY_API_KEY"

# Get analytics

External Transmission

Medium
Category
Data Exfiltration
Confidence
91% confidence
Finding

Configuring a webhook transmits a destination URL and causes the external service to push future event data, including call and lead information, to another endpoint. This creates an ongoing data egress path that can leak sensitive information if misconfigured or set to an untrusted destination.

Content

Scanner excerpt · SKILL.md (reported line 171)May include surrounding context.

Configure a webhook URL to receive real-time events:

bash
curl -s -X POST "https://api.agentizzy.com/api/v1/agents/AGENT_ID/webhook" \
  -H "Authorization: Bearer $AGENTIZZY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://your-endpoint.com/webhook"}'

External Transmission

Medium
Category
Data Exfiltration
Confidence
87% confidence
Finding

Provisioning a dedicated phone number is a state-changing external action that allocates resources and may incur cost or alter how customer communications are routed. In an autonomous agent setting, such side effects are sensitive and should not occur without deliberate authorization.

Content

Scanner excerpt · SKILL.md (reported line 196)May include surrounding context.

Provision a dedicated number (callers reach the agent directly, no access code):

bash
curl -s -X POST "https://api.agentizzy.com/api/v1/agents/AGENT_ID/phone/provision" \
  -H "Authorization: Bearer $AGENTIZZY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"phoneNumber": "+14045551234"}'

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

Including billing usage and checkout operations expands the skill from phone-agent management into financial/account-management actions. In an agentic context, this broadens the blast radius from data retrieval to potentially charge-incurring operations, which is materially more sensitive than the core stated purpose.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
86% confidence
Finding

Billing usage and checkout endpoints transmit account-linked information to a third-party service and may expose commercial data or initiate financial workflows. In the context of an agent skill, these external calls are more sensitive than ordinary analytics queries because they affect billing posture and spending visibility.

Content

Scanner excerpt · SKILL.md (reported line 214)May include surrounding context.

Check current usage and plan details:

bash
curl -s https://api.agentizzy.com/api/v1/billing/usage \
  -H "Authorization: Bearer $AGENTIZZY_API_KEY"

External Transmission

Medium
Category
Data Exfiltration
Confidence
86% confidence
Finding

Billing usage and checkout endpoints transmit account-linked information to a third-party service and may expose commercial data or initiate financial workflows. In the context of an agent skill, these external calls are more sensitive than ordinary analytics queries because they affect billing posture and spending visibility.

Content

Scanner excerpt · SKILL.md (reported line 214)May include surrounding context.

Check current usage and plan details:

bash
curl -s https://api.agentizzy.com/api/v1/billing/usage \
  -H "Authorization: Bearer $AGENTIZZY_API_KEY"

External Transmission

Medium
Category
Data Exfiltration
Confidence
93% confidence
Finding

Posting to the checkout endpoint can initiate a paid-plan upgrade workflow, making this an external state-changing action with financial impact. In an agent skill, that is notably dangerous because a mistaken or automated invocation could lead to unauthorized purchases or billing changes.

Content

Scanner excerpt · SKILL.md (reported line 220)May include surrounding context.

Upgrade to a paid plan:

bash
curl -s -X POST https://api.agentizzy.com/api/v1/billing/checkout \
  -H "Authorization: Bearer $AGENTIZZY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"tier": "builder"}'

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 154)May include surrounding context.

md
exit 1
fi

curl -s "https://api.agentizzy.com/api/v1/agents/${FACILITY_ID}/leads?limit=${LIMIT}" \
  -H "Authorization: Bearer ${AGENTIZZY_API_KEY}"

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 162)May include surrounding context.

md
exit 1
fi

curl -s "https://api.agentizzy.com/api/v1/agents/${FACILITY_ID}/leads?limit=${LIMIT}" \
  -H "Authorization: Bearer ${AGENTIZZY_API_KEY}"

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 187)May include surrounding context.

md
exit 1
fi

curl -s "https://api.agentizzy.com/api/v1/agents/${FACILITY_ID}/leads?limit=${LIMIT}" \
  -H "Authorization: Bearer ${AGENTIZZY_API_KEY}"

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/calls.sh (reported line 21)May include surrounding context.

sh
exit 1
fi

curl -s "https://api.agentizzy.com/api/v1/agents/${FACILITY_ID}/leads?limit=${LIMIT}" \
  -H "Authorization: Bearer ${AGENTIZZY_API_KEY}"

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/leads.sh (reported line 20)May include surrounding context.

sh
exit 1
fi

curl -s "https://api.agentizzy.com/api/v1/agents/${FACILITY_ID}/leads?limit=${LIMIT}" \
  -H "Authorization: Bearer ${AGENTIZZY_API_KEY}"

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/register.sh (reported line 22)May include surrounding context.

sh
exit 1
fi

curl -s "https://api.agentizzy.com/api/v1/agents/${FACILITY_ID}/leads?limit=${LIMIT}" \
  -H "Authorization: Bearer ${AGENTIZZY_API_KEY}"

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/setup.sh (reported line 32)May include surrounding context.

sh
exit 1
fi

curl -s "https://api.agentizzy.com/api/v1/agents/${FACILITY_ID}/leads?limit=${LIMIT}" \
  -H "Authorization: Bearer ${AGENTIZZY_API_KEY}"

External Transmission

Medium
Category
Data Exfiltration
Confidence
70% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/register.sh (reported line 22)May include surrounding context.

sh
[ -n "$AGENT_ID" ] && BODY="${BODY}, \"agent_id\": \"${AGENT_ID}\""
BODY="${BODY}}"

curl -s -X POST "https://api.agentizzy.com/api/v1/register" \
  -H "Content-Type: application/json" \
  -d "$BODY"

External Transmission

Medium
Category
Data Exfiltration
Confidence
70% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/setup.sh (reported line 32)May include surrounding context.

sh
[ -n "$WEBSITE" ] && BODY="${BODY}, \"website\": \"${WEBSITE}\""
BODY="${BODY}}"

curl -s -X POST "https://api.agentizzy.com/api/v1/agents" \
  -H "Authorization: Bearer ${AGENTIZZY_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "$BODY"

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
76% confidence
Finding

The manifest and primary description focus on answering inbound calls, capturing leads, booking appointments, and returning call data. The webhook events section additionally advertises sms.received, which introduces SMS-channel handling not clearly declared in the stated scope.

Content

No source excerpt is available for this finding.