Security audit
Security checks for vulnerabilities and agentic risk
This is a Chinese Dota 2 guide skill with manual data-update scripts, and I found no hidden persistence, credential access, or exfiltration behavior.
Install only if you want a Chinese Dota 2 coaching skill. Use the update scripts knowingly: they may contact public game-data APIs and overwrite the skill's local JSON databases, and a few older utilities assume a specific OpenClaw filesystem path.
The skill description, triggers, and usage examples are entirely in Chinese, and there is no indication that users may interact in another language or that the Chinese-only scope is an intentional, documented locale constraint. This can violate language/locale policy because it implicitly restricts usage to a specific language without user opt-in.
The trigger list includes several broad terms such as “dota”, “dota2”, and generic gameplay-help phrases that can fire on ordinary conversation, increasing the chance of unintended activation. In an agent context, overbroad invocation can cause the wrong skill to respond, leak irrelevant local data, or interfere with user intent even though the content itself is non-privileged game guidance.
The trigger phrases include very broad, common terms such as 'dota' and 'dota2', which are likely to appear in ordinary conversation and can cause accidental invocation of the skill. This increases the chance the assistant enters the skill’s rigid output mode unexpectedly, potentially overriding normal conversational behavior and causing prompt-routing confusion.
The JSON manifest is clearly locale-specific (abilities_cn.json) and most name_loc values are fixed Chinese strings, which can violate a language/locale policy when no user opt-in or alternative selection mechanism is documented in the file. Because this file is a manifest and contains no accompanying natural-language explanation of locale constraints, it presents a forced-language behavior at the configuration level.
The module docstring is written entirely in Chinese and provides the primary human-facing description of the script's behavior. Under the policy, forcing a specific language without user opt-in or a documented justification is a natural-language locale violation.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
import subprocess
def fetch_ability_details():
url = "https://api.opendota.com/api/constants/abilities"
result = subprocess.run(['curl', '-s', url], capture_output=True, text=True)
return json.loads(result.stdout)
This code performs a direct write to the persistent data file heroes_db.json, replacing its contents in place. There is no confirmation prompt, warning message, or explanatory logging before the destructive file write, so users are not clearly informed that local skill data will be modified.
This code file contains natural-language strings that force a specific language/locale for the skill experience, including the module docstring and later CLI help/output text. Under the policy, locale-specific behavior should either be optional or clearly justified; here there is no user opt-in or documented region-specific limitation.
The module docstring says the script fetches each hero's item pick-rate data and saves it to JSON, but the implementation does so by spawning an external command rather than using a direct library call. For an unknown-purpose skill, subprocess execution is a powerful capability not justified by the stated documentation, because it enables behavior beyond simple data retrieval.
This code performs a file write to a fixed path, replacing the existing abilities database contents. The script has no confirmation prompt, warning log before the write, or comment/docstring explicitly disclosing that it will overwrite the target file, so users may not realize an existing dataset will be replaced.
The module docstring presents the script as a merge utility combining fields from two sources. However, OUTPUT_FILE is set to the same path as LOCAL_FILE, so running the script destructively replaces the local source file rather than merely generating a merged artifact from it. This is an intent-level contradiction because the documentation frames the local file as an input provider.
The script writes directly to OUTPUT_FILE, which is the same path as the local items database, replacing its contents in place. Although the file purpose is to merge item data, there is no confirmation prompt, backup step, or explicit user-facing warning in the script before performing this destructive overwrite.
No suspicious patterns detected.