Back to skill

Security audit

Kuaidi100 Package Tracker

Security checks for vulnerabilities and agentic risk

Overview

The skill does what it claims, but its public webhook can accept weakly authenticated package updates that may change local package state and trigger Google Calendar changes.

Review this before installing. Use a high-entropy webhook token, set webhook.signatureMode to strict, configure a callback salt, and avoid exposing the webhook publicly unless needed. Be aware that package numbers, notes, delivery status, and latest tracking context may be stored locally and sent to Kuaidi100 or Google Calendar as part of normal operation.

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 (3)

T09 · Insecure Skill Coding Practices

Error
Location
tracker_core.py:169
Finding

Webhook callbacks are accepted without valid signatures by default

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
index.ts:88
Finding

Unbounded webhook body buffering permits memory-exhaustion denial of service

Content
View full analysis
{ return new Promise((resolve, reject) => { let data = ""; req.on("data", (chunk: Buffer) => { data += chunk.toString(); }); req.on("end", () => resolve(data)); req.on("error", reject); }); } ``` ```typescript const body = await readBody(req); const result = await callCore(cfg(), "handle_push", { push_body: body }) as any; ``` ### Technical Analysis The webhook handler concatenates every received chunk into a JavaScript string without enforcing a maximum byte count. This allows a remote client to make the Node.js process allocate memory proportional to the request size. The request-count rate limiter does not address this issue because a single request can contain an arbitrarily large body. Moreover, the implementation notes that clients behind a tunnel can share the same observed proxy address, making the limiter a coarse control rather than a reliable per-client defense. After buffering, the body is serialized again into the Python subprocess arguments, introducing additional memory and CPU overhead. Extremely large arguments may also exceed operating-system command-line limits and generate repeated subprocess failures. ### Attack Path 1. An attacker reaches the public webhook route. 2. The attacker opens one or more HTTP POST requests. 3. Each request streams a very large body or sends data slowly over an extended period. 4. `readBody` continuously appends received chunks without checking cumulative size. 5. The Node.js process consumes increasing memory and CPU. 6. If the request completes, the oversized body is also serialized and passed toward a Python subprocess, causing additional resource consumption or an argument-size failure. 7. Concurrent requests can exhaust availa ...[truncated 490 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Note
Location
tracker_core.py:57
Finding

Package tracking state is stored without explicitly restrictive file permissions

Content
View full analysis
dict: if STATE_FILE.exists(): try: return json.loads(STATE_FILE.read_text()) except Exception: pass return {"packages": {}} def save_state(state: dict) -> None: STATE_FILE.parent.mkdir(parents=True, exist_ok=True) STATE_FILE.write_text(json.dumps(state, ensure_ascii=False, indent=2)) ``` ### Technical Analysis The state file contains tracking numbers, user notes, carrier information, timestamps, delivery status, and latest tracking context. The code creates the parent directory and writes the file without explicitly setting restrictive permission modes. Actual permissions therefore depend on the process umask and any pre-existing directory or file permissions. On a system with a permissive umask, or where the configured state path points into a shared directory, other local users may be able to read or alter this information. The write is also non-atomic. A crash during `write_text` can leave a partial file, after which `load_state` silently returns an empty package state. ### Attack Path 1. The OpenClaw process runs with a permissive umask or uses a state path in a directory accessible to other local users. 2. The Skill creates or updates `package-tracker-state.json` without enforcing owner-only permissions. 3. Another local account reads the file to obtain tracking and delivery metadata. 4. If write access is also available, that account modifies package state or replaces the file. 5. Subsequent Skill operations consume the modified state and may expose incorrect package information or use it during Calendar synchronization. ### Im ...[truncated 433 chars]
Remediation
View remediation
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
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (29)

Tainted flow: 'req' from os.environ.get (line 316, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

The code sends tracking numbers and callback configuration to Kuaidi100, an external service, as part of subscription setup. In this skill's context that may be functional, but it is still a privacy-relevant external transmission and there is no evidence in this file of consent or data-minimization around sharing package identifiers and webhook details.

Content

Scanner excerpt · dist/tracker_core.py (reported line 135)May include surrounding context.

python
headers={"Content-Type": "application/x-www-form-urlencoded; charset=UTF-8"},
            method="POST",
        )
        with urllib.request.urlopen(req, timeout=20) as resp:
            return json.loads(resp.read().decode())
    except Exception as e:
        return {"error": str(e)}

Tainted flow: 'req' from os.environ.get (line 316, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

The code automatically uploads package details, including notes and delivery context, into Google Calendar whenever a shipment is inferred to be out for delivery. This creates a real privacy risk because potentially sensitive shipment metadata is transmitted to a third party and persisted in calendar entries without any confirmation gate in this component.

Content

Scanner excerpt · dist/tracker_core.py (reported line 319)May include surrounding context.

python
req = urllib.request.Request(url, data=json.dumps(body).encode(), headers=headers, method="POST")
        action = "created"

    with urllib.request.urlopen(req, timeout=15) as resp:
        result = json.loads(resp.read())
        return {"action": action, "event_id": result.get("id"), "summary": body["summary"]}

Tainted flow: 'req' from os.environ.get (line 316, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

The Kuaidi100 subscription request is sent to an endpoint fully configurable by the KUAIDI100_SUB_ENDPOINT environment variable, and the request body includes sensitive API material and webhook registration data. If that environment variable is tampered with, the skill can exfiltrate credentials or send privileged subscription requests to an attacker-controlled host, making this more dangerous because the code handles third-party API secrets.

Content

Scanner excerpt · tracker_core.py (reported line 135)May include surrounding context.

python
headers={"Content-Type": "application/x-www-form-urlencoded; charset=UTF-8"},
            method="POST",
        )
        with urllib.request.urlopen(req, timeout=20) as resp:
            return json.loads(resp.read().decode())
    except Exception as e:
        return {"error": str(e)}

Tainted flow: 'req' from os.environ.get (line 316, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · dist/tracker_core.py (reported line 250)May include surrounding context.

python
headers={"Content-Type": "application/x-www-form-urlencoded"},
        method="POST",
    )
    with urllib.request.urlopen(req, timeout=15) as resp:
        return json.loads(resp.read())["access_token"]

Tainted flow: 'req' from os.environ.get (line 316, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · dist/tracker_core.py (reported line 266)May include surrounding context.

python
headers={"Content-Type": "application/x-www-form-urlencoded"},
        method="POST",
    )
    with urllib.request.urlopen(req, timeout=15) as resp:
        return json.loads(resp.read())["access_token"]

Tainted flow: 'req' from os.environ.get (line 316, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · tracker_core.py (reported line 250)May include surrounding context.

python
headers={"Content-Type": "application/x-www-form-urlencoded"},
        method="POST",
    )
    with urllib.request.urlopen(req, timeout=15) as resp:
        return json.loads(resp.read())["access_token"]

Tainted flow: 'req' from os.environ.get (line 316, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · tracker_core.py (reported line 266)May include surrounding context.

python
headers={"Content-Type": "application/x-www-form-urlencoded"},
        method="POST",
    )
    with urllib.request.urlopen(req, timeout=15) as resp:
        return json.loads(resp.read())["access_token"]

Tainted flow: 'req' from os.environ.get (line 316, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · tracker_core.py (reported line 319)May include surrounding context.

python
headers={"Content-Type": "application/x-www-form-urlencoded"},
        method="POST",
    )
    with urllib.request.urlopen(req, timeout=15) as resp:
        return json.loads(resp.read())["access_token"]

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The README explains subscription to Kuaidi100 push updates and webhook processing, but it does not plainly disclose that tracking numbers and shipment status data are sent to Kuaidi100 and received through a public webhook workflow. This can expose personal shipment metadata to external services and infrastructure, which is especially relevant because package tracking numbers and delivery states can reveal user habits, location patterns, and expected deliveries.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The README explains subscription to Kuaidi100 push updates and webhook processing, but it does not plainly disclose that tracking numbers and shipment status data are sent to Kuaidi100 and received through a public webhook workflow. This can expose personal shipment metadata to external services and infrastructure, which is especially relevant because package tracking numbers and delivery states can reveal user habits, location patterns, and expected deliveries.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
80% confidence
Finding

The skill defines and uses a local state file for package-tracker data, which implies file writes to the user's workspace state area. In this JavaScript file, there is no confirmation prompt or direct user-facing warning that package data will be stored locally.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

Webhook handling automatically calls the Python core to process push data and may sync delivery reminders to Google Calendar, transmitting package-related data to an external service. Although the plugin description mentions auto-adding reminders, this code path provides no explicit user-facing warning at the time of automatic transmission.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

Several tool descriptions specify only Chinese example utterances such as 'Use when user says: 给我加个快递单号' and similar phrases. This creates a natural-language locale bias without explicit user opt-in or a documented justification that the skill is restricted to Chinese-language interactions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The list_packages tool uses only Chinese example requests for invocation guidance, reinforcing a Chinese-only interaction model. Without explicit opt-in or documented regional scope, this conflicts with language/locale neutrality expectations.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The sync_to_calendar tool description provides only Chinese trigger examples. This indicates a forced locale in natural-language guidance without user choice or an explicit, justified locale constraint.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The remove_tracking_number tool description again restricts example user phrasing to Chinese only. This is a natural-language locale policy concern unless the skill explicitly documents a Chinese-only scope.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The skill persists package data locally, including tracking numbers, notes, delivery status, and raw status text, without any disclosure in this file. While local persistence is common, undisclosed storage of potentially sensitive shipment information increases privacy risk, especially in shared environments or when the default path is under a user home directory.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This code transmits package metadata and user notes to Google Calendar, which is a third-party service, without any visible user warning or approval flow in this file. Because notes may contain arbitrary personal information, the privacy impact is higher than a generic calendar sync.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

Several tool descriptions specify activation examples only in Chinese, such as 'Use when user says' followed exclusively by Chinese phrases. This creates a natural-language policy concern because the skill appears to assume or require a specific language without explicit user opt-in or a documented locale limitation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The manifest text says the skill tracks "Chinese domestic packages" and the schedule configuration defaults the timezone to "Asia/Shanghai" without presenting this as a user choice. Under the policy rule, forcing a specific locale is a natural-language policy concern unless it is clearly documented as a justified region-specific tool or offered as opt-in.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The plugin exposes a public webhook and documents a default signature mode of "soft," which effectively allows callbacks to be accepted without strict signature validation. For an internet-reachable endpoint that can influence package state and trigger downstream actions like calendar reminder creation, this weak authentication increases the risk of spoofed requests, event injection, and abuse of the exposed callback surface.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The code defaults TIMEZONE_NAME to Asia/Shanghai, imposing a specific locale setting when the user has not explicitly chosen one. This matches the policy category for language/locale constraints because the skill behavior is region-specific by default rather than opt-in.

Content

No source excerpt is available for this finding.

Tainted flow: 'STATE_FILE' from os.environ.get (line 31, credential/environment) → pathlib.Path.write_text (file write)

Medium
Category
Data Flow
Confidence
65% confidence
Finding

Data from a source is assigned to a variable that is later passed to a sink, creating a variable-mediated taint flow.

Content

Scanner excerpt · dist/tracker_core.py (reported line 69)May include surrounding context.

python
def save_state(state: dict) -> None:
    STATE_FILE.parent.mkdir(parents=True, exist_ok=True)
    STATE_FILE.write_text(json.dumps(state, ensure_ascii=False, indent=2))


# ── Kuaidi100 subscribe API ────────────────────────────────────────────────────

Tainted flow: 'STATE_FILE' from os.environ.get (line 31, credential/environment) → pathlib.Path.write_text (file write)

Medium
Category
Data Flow
Confidence
65% confidence
Finding

Data from a source is assigned to a variable that is later passed to a sink, creating a variable-mediated taint flow.

Content

Scanner excerpt · tracker_core.py (reported line 69)May include surrounding context.

python
def save_state(state: dict) -> None:
    STATE_FILE.parent.mkdir(parents=True, exist_ok=True)
    STATE_FILE.write_text(json.dumps(state, ensure_ascii=False, indent=2))


# ── Kuaidi100 subscribe API ────────────────────────────────────────────────────

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 · dist/tracker_core.py (reported line 73)May include surrounding context.

python
# ── Kuaidi100 subscribe API ────────────────────────────────────────────────────
# Docs: https://api.kuaidi100.com/document/subscribeApi
# Sign = MD5(param + key + customer).upper()
# Endpoint: POST https://poll.kuaidi100.com/poll
# Form fields: schema=json, param=<json>, sign=<sign>, key=<customer>