Back to skill

Security audit

nokey-vehicle-info

Security checks for vulnerabilities and agentic risk

Overview

The skill has a coherent vehicle-status purpose, but it handles sensitive vehicle tokens and location data in ways users should review carefully before installing.

Review this skill before installing. It may expose precise vehicle location and status, stores reusable tokens in plaintext, prints tokens in examples, and contains realistic full token values. Only use it with tokens you can rotate or revoke, avoid sharing transcripts containing credentials, and prefer fixing the documentation to mask secrets, secure the cache, and require explicit confirmation before lookups.

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

Error
Location
SKILL.md:202
Finding

Plaintext Credential Storage and Disclosure in Command Output

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 202-233
Vulnerability Type: Plaintext sensitive-data storage and credential disclosure
Risk Level: High

Vulnerable Code

bash
if [ -f "$CACHE_FILE" ]; then
  accessToken=$(jq -r '.accessToken // empty' "$CACHE_FILE")
  vehicleToken=$(jq -r '.vehicleToken // empty' "$CACHE_FILE")
  
  if [ -n "$accessToken" ] && [ -n "$vehicleToken" ]; then
    echo "认证信息已存在"
    echo "accessToken: $accessToken"
    echo "vehicleToken: $vehicleToken"
  else
    echo "认证信息不完整,请提供 vehicleToken####accessToken"
  fi
else
  echo "未找到认证信息,请提供 vehicleToken####accessToken"
fi
bash
# 解析用户输入的 token(格式为 vehicleToken####accessToken)
AUTH_INPUT="xxxxxx####xxxxx"
vehicleToken=$(echo "$AUTH_INPUT" | awk -F '####' '{print $1}')
accessToken=$(echo "$AUTH_INPUT" | awk -F '####' '{print $2}')

# 保存到缓存文件(默认使用 PROD 环境)
cat > ~/.nokey_vehicle_info_cache.json << EOF
{
  "accessToken": "$accessToken",
  "vehicleToken": "$vehicleToken",
  "env": "PROD",
  "last_updated": "$(date -Iseconds)"
}
EOF

Technical Analysis

The documented implementation prints both authentication tokens directly to standard output and stores them in an unencrypted JSON file. The cache creation does not set a restrictive umask or explicitly apply mode 0600. Consequently, effective permissions depend on the user's environment and may allow other local accounts or processes to read the file.

Printing the values can additionally expose them through terminal scrollback, shell automation logs, agent transcripts, CI logs, monitoring systems, or support bundles. These are bearer-style credentials, so possession may be sufficient to issue requests without further proof of identity.

The file contains both credentials required by the API:

  • accessToken, transmitted in the Authorization header.
  • vehicleToken, transmitted in the request ...[truncated 1270 chars]
Remediation
View remediation

Remediation Suggestions

  1. Remove all commands that print complete token values. Display only a fixed confirmation or a minimally redacted fingerprint.
  2. Avoid collecting credentials through persistent conversation history where possible. Use a dedicated secret-input or credential-management mechanism.
  3. Store credentials in an operating-system keychain, secret service, or narrowly scoped credential store rather than a plaintext JSON file.
  4. If a local file is unavoidable, create it with restrictive permissions before writing:
    bash
    CACHE_FILE="$HOME/.nokey_vehicle_info_cache.json"
    umask 077
    tmp=$(mktemp "${CACHE_FILE}.XXXXXX") || exit 1
    chmod 600 "$tmp" || exit 1
    
  5. Write validated JSON to the temporary file and atomically rename it to the destination.
  6. Reject symbolic links and verify that any existing cache is a regular file owned by the current user.
  7. Implement token expiration, rotation, deletion, and revocation procedures.
  8. Ensure logs, errors, API responses, and diagnostic output redact authorization material.

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:375
Finding

Complete Token-Shaped Credentials Embedded in Documentation

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 375-383; README.md, lines 178-184
Vulnerability Type: Hardcoded secrets or sensitive credential material
Risk Level: High

Vulnerable Code

SKILL.md includes the following cache example:

json
{
  "vehicleToken": "sk-api-R44YuYJmggYqQi3nMch1uY8jBhJEToNmAZyMu6Z4nvlncdaDmqfGRtqJ2LWcvo63mGRsgELugjvnAU1QXzp5K7NKdxaLabXqd5yQBolsTeQpCw6rUsNH7fPRIcYE8tUc",
  "accessToken": "eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJsb2dpblR5cGUiOiJsb2dpbiIsImxvZ2luSWQiOjIzLCJyblN0ciI6IlBtNDhXUmV0SlAzczNLb1R4UjNpNWg0MHRLbHpHNExsIiwidGVuYW50X2NvZGUiOiJDSEVOR1FVIiwidGVuYW50X2lkIjoxLCJhcHBfaWQiOiJnNml0eWx4eHg1a2FiajFmIn0.HkbKnHimmyfY0b_ZGNBXyzVOpb0KxQhV58roMc84ueQ",
  "env": "PROD",
  "last_updated": "2026-03-20T16:57:53+08:00"
}

README.md contains another complete JWT-shaped value:

json
{
  "accessToken": "eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJsb2dpblR5cGUiOiJsb2dpbiIsImxvZ2luSWQiOjIzLCJyblN0ciI6InhUc3RlaGk2ZDJmQ3BiN0tVWk5aeU5yb0t0MTVLSzk5IiwidGVuYW50X2NvZGUiOiJDSEVOR1FVIiwidGVuYW50X2lkIjoxLCJhcHBfaWQiOiJnNml0eWx4eHg1a2FiajFmIn0.r5phB61twZzR3NunoWeLBEakarqHCCeWB-VtcXrliXc",
  "vehicleToken": "5c32cc588a8ef984a02610a59211616c",
  "last_updated": "2026-03-20T16:57:53+08:00"
}

Technical Analysis

The documentation publishes complete values shaped like the credentials accepted by the vehicle API. The access-token values have the three-part structure of JSON Web Tokens, while the accompanying vehicle-token values are presented in the exact cache fields consumed by the documented commands.

Static inspection cannot prove that these values are currently valid. Nevertheless, complete production-like credentials must be treated as potentially compromised because source repositories, package registries, mirrors, forks, caches, and historical commits can retain them after removal.

Attack Path

  1. An attacker obtains the project from a repositor ...[truncated 1083 chars]
Remediation
View remediation

Remediation Suggestions

  1. Immediately verify whether all documented tokens were ever issued by the service.
  2. Revoke and rotate every potentially real token, including values appearing in earlier revisions or distributed packages.
  3. Replace complete values with unmistakably invalid placeholders:
    json
    {
      "accessToken": "<REDACTED_ACCESS_TOKEN>",
      "vehicleToken": "<REDACTED_VEHICLE_TOKEN>"
    }
    
  4. Remove the values from version-control history and published artifacts where operationally possible. Rotation remains mandatory because history rewriting does not revoke copied credentials.
  5. Add automated secret scanning to pre-commit checks and continuous integration.
  6. Maintain examples in sanitized fixtures that cannot authenticate to any environment.

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:338
Finding

Predictable Shared Temporary File Enables Symlink-Based File Overwrite

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 338-343
Vulnerability Type: Insecure predictable temporary file
Risk Level: Medium

Vulnerable Code

bash
# 切换到 UAT 环境
jq '.env = "UAT"' ~/.nokey_vehicle_info_cache.json > /tmp/cache.json && mv /tmp/cache.json ~/.nokey_vehicle_info_cache.json

# 切换到 PROD 环境
jq '.env = "PROD"' ~/.nokey_vehicle_info_cache.json > /tmp/cache.json && mv /tmp/cache.json ~/.nokey_vehicle_info_cache.json

Technical Analysis

Both environment-switch commands write to the fixed path /tmp/cache.json. On typical multi-user systems, /tmp is shared and permits users to create files and symbolic links. Shell redirection opens the destination before jq runs and follows symbolic links.

An attacker who can create /tmp/cache.json first can point it to another file writable by the victim. When the victim executes the command, the shell truncates and overwrites the linked target with JSON output. The subsequent mv can also produce unexpected behavior depending on ownership and sticky-directory protections.

The attack requires local access and timing or prior placement of the predictable path. It does not by itself bypass the victim user's operating-system permissions.

Attack Path

  1. A local attacker predicts that the victim will use the documented environment-switch command.
  2. The attacker creates /tmp/cache.json as a symbolic link to a file that the victim can write.
  3. The victim executes the command to switch between UAT and production.
  4. Shell redirection follows the symbolic link and truncates the target file.
  5. jq writes the modified cache JSON into that target, corrupting its prior contents.
  6. Depending on the target selected, the corruption can disrupt user configuration, application operation, or future shell behavior.

Impact Assessment

The attacker can cause overwrite or corruption of files writable by the u ...[truncated 383 chars]

Remediation
View remediation

Remediation Suggestions

Use a unique temporary file in a user-controlled directory, apply restrictive permissions, and rename it atomically only after successful validation:

bash
CACHE_FILE="$HOME/.nokey_vehicle_info_cache.json"
umask 077
tmp=$(mktemp "${CACHE_FILE}.XXXXXX") || exit 1
trap 'rm -f "$tmp"' EXIT

jq '.env = "UAT"' "$CACHE_FILE" > "$tmp" || exit 1
chmod 600 "$tmp" || exit 1
mv -- "$tmp" "$CACHE_FILE" || exit 1
trap - EXIT

Additional hardening should include:

  1. Confirm that the cache is a regular file owned by the current user.
  2. Reject symbolic-link cache destinations.
  3. Keep temporary credential data out of globally shared directories.
  4. Validate the generated JSON before replacing the existing cache.
  5. Preserve a secure backup or use locking if concurrent invocations are possible.

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:33
Finding

Conflicting Credential Formats Can Swap Tokens Across Protocol Fields

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 33-42 and 220-224; README.md, lines 46, 70, and 195
Vulnerability Type: Ambiguous security-sensitive input parsing
Risk Level: Medium

Vulnerable Code

SKILL.md defines and parses the format as vehicleToken####accessToken:

text
vehicleToken####accessToken
bash
# 解析用户输入的 token(格式为 vehicleToken####accessToken)
AUTH_INPUT="xxxxxx####xxxxx"
vehicleToken=$(echo "$AUTH_INPUT" | awk -F '####' '{print $1}')
accessToken=$(echo "$AUTH_INPUT" | awk -F '####' '{print $2}')

In contrast, README.md tells users to submit the reverse order:

text
Token format: accessToken####vehicleToken

The resulting variables are transmitted in different security-sensitive protocol fields:

bash
curl -s -X POST "https://openapi.nokeeu.com/iot/v1/condition" \
  -H "Authorization: Bearer ${accessToken}" \
  -H "Content-Type: application/json" \
  -d "{\"vehicleToken\": \"${vehicleToken}\"}" | jq .

Technical Analysis

The two documents define opposite ordering for the same delimiter-based credential input. The parser shown in SKILL.md always treats the first field as vehicleToken and the second as accessToken, without validating either token type.

A user following README.md will therefore cause the access token to be stored as vehicleToken and sent in the JSON request body. The vehicle token will be stored as accessToken and sent in the Authorization header.

Although both values are sent to the same documented API host, they enter different server-side processing and logging paths. Request bodies and malformed authorization headers may have different retention, monitoring, and diagnostic policies. This creates avoidable credential exposure and predictable authentication failure.

Attack Path

  1. A user follows README.md and submits accessToken####vehicleToken.
  2. The implementation follows `SK ...[truncated 961 chars]
Remediation
View remediation

Remediation Suggestions

  1. Establish one canonical credential schema and update every document and example consistently.
  2. Prefer named structured fields over positional delimiter parsing:
    json
    {
      "accessToken": "<ACCESS_TOKEN>",
      "vehicleToken": "<VEHICLE_TOKEN>"
    }
    
  3. If interactive collection is required, request each value separately through a secure secret-input mechanism.
  4. Validate the expected token structure, allowed length, character set, and number of input fields before storage.
  5. Reject inputs containing missing or additional delimiter-separated fields.
  6. Confirm the parsed field names without echoing their values.
  7. Add automated tests verifying that the access token is used only in the authorization header and the vehicle token only in the intended request-body property.
  8. Clear any previously generated cache because it may already contain swapped credentials.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (22)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
85% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · README.md (reported line 189)May include surrounding context.

如需清除缓存,删除该文件即可:

bash
rm ~/.nokey_vehicle_info_cache.json

常见问题

YARA rule 'agent_skill_mcp_tool_poisoning_metadata': MCP/tool metadata poisoning indicators in tool schemas or skill manifests [agent_skills]

High
Category
YARA Match
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.

md
---
name: nokey-vehicle-info
description: 车辆信息查询技能,支持查询车辆位置、车况信息(车锁、车门、车窗、空调、引擎状态等)。当用户查询车辆位置、询问车辆在哪里、查询车况信息时自动调用此技能。
---

## 何时使用

在以下场景中应用本技能:

- 用户查询车辆位置信息
- 用户询问"我的车在哪儿"或"车辆在哪里"
- 用户请求查询车况信息(电压、电量、里程、续航等)
- 用户询问车辆状态(车锁、车门、车窗、空调是否开启)
- 用户请求获取车辆详细信息(SN、VIN、档位�

Missing User Warnings

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill instructs storing high-sensitivity access tokens and vehicle tokens in a local plaintext cache file under the home directory, but does not prominently warn about the confidentiality and persistence risks. Because these tokens appear sufficient to query sensitive vehicle information, compromise of the local file can enable unauthorized tracking and status access.

Content

No source excerpt is available for this finding.

Ssd 3

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill examples instruct the assistant to reveal cached tokens back in normal output, effectively turning stored secrets into displayable content. In an agent setting, this is especially dangerous because outputs may be logged, shown to end users, or captured by downstream systems, enabling credential theft and later vehicle-data access.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 244)May include surrounding context.

md
# 从缓存读取 token 和环境配置
accessToken=$(jq -r '.accessToken' ~/.nokey_vehicle_info_cache.json)
vehicleToken=$(jq -r '.vehicleToken' ~/.nokey_vehicle_info_cache.json)
ENV=$(jq -r '.env' ~/.nokey_vehicle_info_cache.json)

# 选择环境(缓存中的 env 为 UAT 时使用 UAT,否则使用 PROD)
if [ "$ENV" = "UAT" ]; then

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 340)May include surrounding context.

bash
# 切换到 UAT 环境
jq '.env = "UAT"' ~/.nokey_vehicle_info_cache.json > /tmp/cache.json && mv /tmp/cache.json ~/.nokey_vehicle_info_cache.json

# 切换到 PROD 环境
jq '.env = "PROD"' ~/.nokey_vehicle_info_cache.json > /tmp/cache.json && mv /tmp/cache.json ~/.nokey_vehicle_info_cache.json

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 343)May include surrounding context.

bash
# 切换到 UAT 环境
jq '.env = "UAT"' ~/.nokey_vehicle_info_cache.json > /tmp/cache.json && mv /tmp/cache.json ~/.nokey_vehicle_info_cache.json

# 切换到 PROD 环境
jq '.env = "PROD"' ~/.nokey_vehicle_info_cache.json > /tmp/cache.json && mv /tmp/cache.json ~/.nokey_vehicle_info_cache.json

Ssd 3

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The documentation includes live-looking full credential examples for both vehicleToken and accessToken, which may be copied, reused, or mistakenly treated as valid secrets. Even if illustrative, publishing realistic full-length credentials trains unsafe handling patterns and risks accidental exposure if the values are genuine.

Content

No source excerpt is available for this finding.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · README.md (reported line 35)May include surrounding context.

md
brew install curl jq

# Linux (Ubuntu/Debian)
sudo apt-get install curl jq

# Linux (CentOS/RHEL)
sudo yum install curl jq

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · README.md (reported line 38)May include surrounding context.

sudo apt-get install curl jq

Linux (CentOS/RHEL)

sudo yum install curl jq

text

## 功能特性

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill documentation explicitly instructs caching accessToken and vehicleToken in a local file under the user's home directory, but does not warn that this is plaintext credential storage or require restrictive file permissions. If another local user, process, backup system, or malware accesses that file, the attacker can reuse the tokens to query vehicle status and location.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The README states the skill should auto-trigger when a user asks about vehicle location or status, but the trigger description is broad and lacks clear authorization or confirmation requirements. In a vehicle-information context, overly permissive invocation can cause unintended disclosure of sensitive location and status data to the wrong user or in the wrong conversational context.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
95% confidence
Finding

The documented curl command sends sensitive authentication material and a vehicle identifier to an external API endpoint, which is expected for the skill's function but still represents external transmission of secrets and sensitive telemetry-related data. In this context, disclosure risk is elevated because the same README also encourages local token caching and direct shell use, increasing the chance of accidental leakage through history, logs, or misuse.

Content

Scanner excerpt · README.md (reported line 95)May include surrounding context.

md
accessToken=$(jq -r '.accessToken' ~/.nokey_vehicle_info_cache.json)
vehicleToken=$(jq -r '.vehicleToken' ~/.nokey_vehicle_info_cache.json)

curl -s -X POST "https://openapi.nokeeu.com/iot/v1/condition" \
  -H "Authorization: Bearer ${accessToken}" \
  -H "Content-Type: application/json" \
  -d "{\"vehicleToken\": \"${vehicleToken}\"}" | jq .

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The manifest says the skill should auto-invoke for broad vehicle-related queries such as asking where the car is or querying vehicle status, without narrow scoping or confirmation. In context, this can cause the agent to trigger network calls and sensitive token-handling behavior on ambiguous everyday prompts, increasing unintended data access and disclosure risk.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The workflow normalizes requesting raw credentials from the user and storing them for future reuse, which expands the lifetime and exposure surface of sensitive authentication material. In this context, reused vehicle and access tokens can grant continued access to location and condition data beyond the immediate session.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The API usage section describes transmitting bearer and vehicle tokens to external endpoints to retrieve location and condition data, but does not clearly foreground the privacy implications of sending authentication material and requesting precise vehicle telemetry. In a vehicle-tracking context, users should be explicitly informed that sensitive credentials and location-related data are sent off-host.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The documented cache-check flow explicitly prints both accessToken and vehicleToken to normal output, which can expose reusable credentials to users, logs, chat transcripts, or other observers. For a vehicle-info skill, revealing raw cached secrets is unnecessary to fulfill the intended function and materially increases the risk of account or vehicle-data compromise.

Content

No source excerpt is available for this finding.

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 254)May include surrounding context.

md
fi

# 调用 API
RESPONSE=$(curl -s -X POST "${BASE_URL}/iot/v1/condition" \
  -H "Authorization: Bearer ${accessToken}" \
  -H "Content-Type: application/json" \
  -d "{\"vehicleToken\": \"${vehicleToken}\"}")

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 292)May include surrounding context.

md
fi

# 调用 API
RESPONSE=$(curl -s -X POST "https://openapi.nokeeu.com/iot/v1/condition" \
  -H "Authorization: Bearer ${accessToken}" \
  -H "Content-Type: application/json" \
  -d "{\"vehicleToken\": \"${vehicleToken}\"}")

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The README presents the skill description and interaction examples entirely in Chinese, and the demonstrated user prompts and assistant responses assume Chinese-language usage without mentioning alternatives. For a general-purpose skill, this can be a language policy issue when no user opt-in or locale-specific justification is provided.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

整个技能描述、示例交互和指令均固定为中文,没有说明是否支持其他语言或允许用户选择语言。若组织要求不应在无用户同意下强制特定语言/地区表达,这种默认中文单语设计构成自然语言策略风险。

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The environment-switching instructions allow broad natural-language triggers like 'use UAT environment' or 'use production environment' to automatically mutate persistent configuration. While not directly exposing credentials, it can redirect future authenticated requests to a different backend without adequate confirmation or scope control.

Content

No source excerpt is available for this finding.