Back to skill

Security audit

heyi-paid-api

Security checks for vulnerabilities and agentic risk

Overview

The skill appears to be a legitimate paid Heyi API integration, but users should review it because its documented client snippets can send the API key to a configurable host and the artifact includes risky maintainer publishing tooling.

Install only if you intend to use Heyi's paid API service. Keep HEYI_API_BASE_URL at the default production host unless you fully trust the alternate endpoint, store HEYI_API_KEY in a secret store or environment variable, and require confirmation before paid or batch calls. Publishers should pin the ClawHub CLI and fix the temporary-file handling before using the included release script with real tokens.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (3)

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:30
Finding

Bearer API Key Can Be Sent to an Untrusted Configurable Host

Content
View full analysis
dict: headers = {"Authorization": f"Bearer {KEY}", "Accept": "application/json"} if json_body is not None: headers["Content-Type"] = "application/json" r = httpx.request(method, f"{BASE}{path}", params=params, json=json_body, headers=headers, timeout=30.0) ``` ### Technical Analysis The supplied Node.js and Python clients read `HEYI_API_BASE_URL` and attach the bearer credential to the resulting destination without validating: - The URL scheme - The destination hostname - The destination port - Whether the request path begins with `/api/external/` - Whether the destination belongs to an explicitly approved staging environment The surrounding instructions state that arbitrary hosts must not be u ...[truncated 1621 chars]
Remediation
View remediation

T08 · Insecure Dependencies

Error
Location
bin/publish-all.sh:226
Finding

Mutable clawhub@latest Package Executes While Receiving a Publishing Token

Content
View full analysis
/dev/null 2>&1 || true ``` ### Technical Analysis The publishing script invokes `npx -y clawhub@latest`. This downloads and executes whichever package version currently owns the `latest` tag. No exact version, lockfile resolution, or package integrity value constrains the executable code. The downloaded process then receives `CLAWHUB_TOKEN` through standard input. Although standard input is preferable to placing a token in command-line arguments, it does not protect the token from the process intentionally reading it. Trusting a mutable package version therefore creates a direct supply-chain credential boundary. The package also runs with the maintainer's user permissions and from within the project environment. A malicious package could access repository contents, user-readable files, environment variables, and network resources in addition to capturing the supplied token. ### Attack Path 1. An attacker compromises the upstream `clawhub` package, its publisher account, its release workflow, or the registry serving the package. 2. The attacker publishes a malicious release and assigns it the `latest` tag. 3. A maintainer runs `bin/publish-all.sh clawhub` or `bin/publish-all.sh all`. 4. `npx -y` retrieves and executes the malicious release without an interactive package confirmation. 5. The publishing script supplies `CLAWHUB_TOKEN` to that process through standard input. 6. The malicious package reads and exfiltrates the token. 7. It may al ...[truncated 707 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
bin/publish-all.sh:130
Finding

Predictable Shared Temporary File Allows Symlink-Based File Overwrite

Content
View full analysis
/dev/null; then ok "SkillHub 发布成功 $VERSION" else err "SkillHub 发布失败,响应:" cat /tmp/skillhub_publish.json >&2 return 1 fi ``` ### Technical Analysis The script writes a network response to the fixed path `/tmp/skillhub_publish.json`. Shared temporary directories are generally writable by other local users. The script does not: - Create the file exclusively - Verify its owner or file type - Reject symbolic links - Use a private temporary directory - Assign a restrictive `umask` - Remove the response file after use A local attacker can pre-create the predictable path as a symbolic link to another file writable by the victim. When `curl -o` opens the path, it can follow the link and overwrite the target with the HTTP response. The fixed name also creates a race between concurrent publishing operations. One process can overwrite the response while another process is parsing it. ### Attack Path 1. The attacker has local access to the same machine and can write to `/tmp`. 2. The attacker predicts that a maintainer will run the SkillHub publication command. 3. The attacker creates `/tmp/skillhub_publish.json` as a symbolic link to a file writable by the maintainer account. 4. The maintainer runs `bin/publish-all.sh skillhub` or the `all` command. 5. `curl -o /tmp/skillhub_publish.json` follows the symbolic link and truncates or overwrites the target file with the server response. 6. ...[truncated 717 chars]
Remediation
View remediation
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (69)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% 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 151)May include surrounding context.

md
npm uninstall -g heyihub-skill

# 2. 删除 Skill 目录
rm -rf ~/.cursor/skills/heyi-paid-api
rm -rf ~/.claude/skills/heyi-paid-api
rm -rf ~/.codex/skills/heyi-paid-api
rm -rf ~/.mavis/skills/heyi-paid-api

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% 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 152)May include surrounding context.

md
npm uninstall -g heyihub-skill

# 2. 删除 Skill 目录
rm -rf ~/.cursor/skills/heyi-paid-api
rm -rf ~/.claude/skills/heyi-paid-api
rm -rf ~/.codex/skills/heyi-paid-api
rm -rf ~/.mavis/skills/heyi-paid-api

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% 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 153)May include surrounding context.

md
npm uninstall -g heyihub-skill

# 2. 删除 Skill 目录
rm -rf ~/.cursor/skills/heyi-paid-api
rm -rf ~/.claude/skills/heyi-paid-api
rm -rf ~/.codex/skills/heyi-paid-api
rm -rf ~/.mavis/skills/heyi-paid-api

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% 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 154)May include surrounding context.

md
npm uninstall -g heyihub-skill

# 2. 删除 Skill 目录
rm -rf ~/.cursor/skills/heyi-paid-api
rm -rf ~/.claude/skills/heyi-paid-api
rm -rf ~/.codex/skills/heyi-paid-api
rm -rf ~/.mavis/skills/heyi-paid-api

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% 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 155)May include surrounding context.

md
npm uninstall -g heyihub-skill

# 2. 删除 Skill 目录
rm -rf ~/.cursor/skills/heyi-paid-api
rm -rf ~/.claude/skills/heyi-paid-api
rm -rf ~/.codex/skills/heyi-paid-api
rm -rf ~/.mavis/skills/heyi-paid-api

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% 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 151)May include surrounding context.

md
npm uninstall -g heyihub-skill

# 2. 删除 Skill 目录
rm -rf ~/.cursor/skills/heyi-paid-api
rm -rf ~/.claude/skills/heyi-paid-api
rm -rf ~/.codex/skills/heyi-paid-api
rm -rf ~/.mavis/skills/heyi-paid-api

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% 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 152)May include surrounding context.

md
# 2. 删除 Skill 目录
rm -rf ~/.cursor/skills/heyi-paid-api
rm -rf ~/.claude/skills/heyi-paid-api
rm -rf ~/.codex/skills/heyi-paid-api
rm -rf ~/.mavis/skills/heyi-paid-api
rm -rf ~/.openclaw/skills/heyi-paid-api

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% 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 153)May include surrounding context.

2. 删除 Skill 目录

rm -rf ~/.cursor/skills/heyi-paid-api rm -rf ~/.claude/skills/heyi-paid-api rm -rf ~/.codex/skills/heyi-paid-api rm -rf ~/.mavis/skills/heyi-paid-api rm -rf ~/.openclaw/skills/heyi-paid-api

text

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% 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 154)May include surrounding context.

rm -rf ~/.cursor/skills/heyi-paid-api rm -rf ~/.claude/skills/heyi-paid-api rm -rf ~/.codex/skills/heyi-paid-api rm -rf ~/.mavis/skills/heyi-paid-api rm -rf ~/.openclaw/skills/heyi-paid-api

text

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% 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 155)May include surrounding context.

rm -rf ~/.claude/skills/heyi-paid-api rm -rf ~/.codex/skills/heyi-paid-api rm -rf ~/.mavis/skills/heyi-paid-api rm -rf ~/.openclaw/skills/heyi-paid-api

text

只删 Skill 目录也可正常工作,但 npm 缓存里的 `snapshots/catalog.json` 会残留;建议两步都做。

Tp2

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names.

Content

No source excerpt is available for this finding.

Tp2

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names.

Content

No source excerpt is available for this finding.

Tp2

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names.

Content

No source excerpt is available for this finding.

Tp2

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names.

Content

No source excerpt is available for this finding.

Tp2

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names.

Content

No source excerpt is available for this finding.

Tp2

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

If the broader publication and registry-management behaviors described by the analyzer are indeed part of the associated package/workflow but omitted from the skill's declared purpose, that is a transparency and privilege-boundary problem. Hidden or under-declared release/publishing capabilities can be dangerous because they may lead agents or users to execute registry-upload or metadata-submission actions under the guise of a normal API integration skill.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If the broader publication and registry-management behaviors described by the analyzer are indeed part of the associated package/workflow but omitted from the skill's declared purpose, that is a transparency and privilege-boundary problem. Hidden or under-declared release/publishing capabilities can be dangerous because they may lead agents or users to execute registry-upload or metadata-submission actions under the guise of a normal API integration skill.

Content

No source excerpt is available for this finding.

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 473)May include surrounding context.

md
A:常见原因是 `HEYI_API_BASE_URL` 指向了旧域名(`api.heyi.cloud` 之类),或 Agent 工作台没重启仍加载旧 `SKILL.md`。先确认 Base URL,再重启 Agent。

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 528)May include surrounding context.

md
:升级 Skill 之后、首次接入、长期未升级、排查 `4040` 之前。开发期可用 `npx heyihub-skill snapshot`(或 `node bin/install.js snapshot`)重新生成 `snapshots/catalog.json`,随 npm release 发布。

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

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

Content

Scanner excerpt · bin/publish-all.sh (reported line 16)May include surrounding context.

sh
#
# 凭据通过环境变量提供(避免写进文件/日志):
#   SKILLHUB_TOKEN          — SkillHub 的 skh_xxx
#   NPM_TOKEN               — npm 的 npm_xxx(或用 ~/.npmrc)
#   GITHUB_TOKEN            — 用于 ClawHub GitHub linking / SkillsMP 自动索引
#   LOBEHUB_TOKEN           — LobeHub API(如支持)

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.