Back to skill

Security audit

Summarize-AI 内容摘要助手

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward summarization helper, but users should understand that files and web content may be sent to external model or extraction services and that it installs a third-party CLI through Homebrew.

Install only if you trust the summarize CLI and its Homebrew tap. Avoid using it on secrets, confidential documents, regulated data, or private pages unless you are comfortable with the selected model provider and optional extraction services receiving that content.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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 (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:5
Finding
Unpinned Executable Installed from a Third-Party Homebrew Tap## Vulnerability Details **File Location**: `SKILL.md`, line 5 **Vulnerability Type**: Unpinned third-party dependency **Risk Level**: Medium ### Vulnerable Code ```yaml metadata: {"clawdbot":{"emoji":"🧾","requires":{"bins":["summarize"]},"install":[{"id":"brew","kind":"brew","formula":"steipete/tap/summarize","bins":["summarize"],"label":"Install summarize (brew)"}]}} ``` ### Technical Analysis The skill instructs the environment to install the `summarize` executable from the third-party Homebrew tap `steipete/tap`. The dependency is not pinned to an immutable version, commit, checksum, or signed artifact. Consequently, the executable retrieved during installation may differ from the component that existed when this skill was audited. This creates a supply-chain trust boundary: control of the tap, its formula, or the referenced release assets could allow an attacker to substitute malicious code. The project does not provide an integrity digest, signature-verification procedure, or other mechanism that would detect such substitution before installation. The use of a third-party tap is not itself evidence that the current package is malicious. The risk arises from mutable external package resolution without integrity pinning. ### Attack Path 1. An attacker compromises the Homebrew tap, its maintainer account, the formula repository, or a mutable release asset referenced by the formula. 2. The attacker modifies the formula or upstream artifact so that it installs an attacker-controlled executable. 3. An agent or user loads the skill and follows its documented Homebrew installation action. 4. Homebrew resolves the unpinned formula and retrieves the modified package content. 5. The malicious executable runs when the skill invokes `summarize`, inheriting the installing user's privileges and accessible environment. ### Impact Assessment Successful exploitation could execute arbitrary code with the privileges of the user ...[truncated 710 chars]
Remediation
## Remediation Suggestions 1. Pin the dependency to an immutable, reviewed version rather than resolving the latest tap formula. 2. Verify downloaded artifacts using a published SHA-256 checksum or a trusted cryptographic signature before execution. 3. Prefer an authenticated official distribution channel with documented release provenance. 4. If the Homebrew tap must remain in use, pin its repository to a reviewed commit and verify that the formula references immutable release assets. 5. Record the expected package version and integrity value in the skill metadata so future audits can reproduce the installed artifact. 6. Run the executable with least privilege and expose only the provider credential required for the selected operation. 7. Periodically review the tap ownership, formula changes, upstream artifact provenance, and signing practices.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill explicitly encourages summarizing URLs, local files, PDFs, images, audio, and YouTube content using third-party model providers, but it does not warn users that the underlying content may be transmitted to external services via provider APIs or fallback services such as Firecrawl and Apify. This creates a real data exposure risk: users may unintentionally send sensitive local documents or private web content off-device under the assumption that summarization is purely local.

Static analysis

No suspicious patterns detected.