Back to skill

Security audit

t0cksn1per

Security checks for vulnerabilities and agentic risk

Overview

This skill is narrowly about running a Tock reservation tool, but it tells the agent to execute an unpinned third-party package that can change after review.

Install only if you are comfortable allowing an agent to fetch and run the current `t0cksn1per` package with your user-level permissions. Prefer a pinned, reviewed version and run it in a constrained environment, especially for unattended node use or CDP/browser automation.

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:18
Finding
Unpinned Third-Party Package Retrieval and Execution## Vulnerability Details **File Location**: `SKILL.md`, lines 18–22 **Vulnerability Type**: Unpinned third-party dependency execution **Risk Level**: Medium ### Vulnerable Code ```markdown ## Command Source Run with: ```bash uvx t0cksn1per --help ``` ``` The same unpinned package is also executed in: - `references/commands.md`, lines 6–11, 16–21, and 25–32 - `references/cdp.md`, lines 14–21 Representative command: ```bash PLAYWRIGHT_HEADLESS=0 uvx t0cksn1per run taneda \ --size 3 \ --release-at 11:00 \ --newly-released-only \ --dates 2026-05-27,2026-05-28,2026-05-29,2026-05-30,2026-05-31 \ --exact-times "5:15 PM,7:45 PM" ``` ### Technical Analysis The Skill directs the Agent to use `uvx` to resolve and execute `t0cksn1per` without specifying an immutable version, validating package hashes, using a lockfile, or otherwise verifying the retrieved artifact. Even the documented `--help` command executes package-controlled Python code. Consequently, the effective executable payload is not fixed by the reviewed project. It can change when a new package version is published or if the package distribution account or package index is compromised. The repository contains no local implementation that would allow the behavior of the executed reservation tool to be verified as part of this audit. This is a supply-chain weakness rather than evidence that the current `t0cksn1per` package is malicious. ### Attack Path 1. An attacker compromises the package publisher, package distribution channel, or a future package release. 2. The attacker introduces malicious initialization or CLI code into the version resolved as `t0cksn1per`. 3. A user invokes the Skill, which instructs the Agent to run `uvx t0cksn1per --help` or `uvx t0cksn1per run ...`. 4. `uvx` retrieves and executes the attacker-controlled package version. 5. The malicious code runs under the invoking user or Agent account before or during the requested reservation operation. 6. It can then acces ...[truncated 817 chars]
Remediation
## Remediation Suggestions 1. Pin `t0cksn1per` to a specifically reviewed version rather than allowing `uvx` to resolve the latest available release. 2. Use an audited dependency lockfile and require cryptographic hash verification for all downloaded artifacts and transitive dependencies. 3. Configure `uv` to use an explicitly trusted package index and prevent fallback to untrusted or unintended indexes. 4. Resolve and verify the package before execution. Do not treat `--help` as safe verification because loading the CLI may execute package-controlled code. 5. Consider maintaining a reviewed wrapper or vendored artifact whose integrity can be checked independently before invocation. 6. Run the dependency with least privilege in an isolated environment, limiting filesystem, environment-variable, browser, and network access to what is required. 7. Retain the existing localhost-only CDP URL and explicit opt-in requirement. When CDP is used, launch Chrome with a dedicated temporary profile containing no unrelated authenticated sessions or sensitive browsing data.
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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
Findings (5)

Rp1

Medium
Category
MCP Rug Pull
Confidence
91% confidence
Finding
The skill instructs the agent to execute `uvx t0cksn1per --help` without pinning a specific package version or source, which causes code to be fetched and run from a mutable upstream at execution time. In an agent skill context, this is more dangerous than ordinary documentation because the workflow explicitly says to generate a command and run it, so a compromised or typosquatted package update could result in arbitrary code execution on the user's local Mac or remote node.

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding
The command invokes `uvx t0cksn1per` without pinning an exact package version, so execution may pull whatever version is currently published. That creates a supply-chain risk: a compromised upstream release or unexpected breaking change could cause arbitrary code execution on the user's Mac or remote node when they follow the skill instructions.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding
The command uses `uvx t0cksn1per` without pinning an exact package version, so execution may fetch and run whatever version is current at invocation time. In a skill that instructs users to launch automation locally or on a remote node, this creates a supply-chain risk: a compromised upstream release or unexpected breaking update could execute unreviewed code in a privileged environment.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding
This example again invokes `uvx t0cksn1per` without a pinned version, meaning the package resolved at runtime is not fixed or reproducible. Because this variant is intended for a headless remote node, exploitation could affect unattended infrastructure and make malicious package changes harder to notice.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding
The CDP launch example also relies on an unpinned `uvx t0cksn1per` package, exposing users to arbitrary upstream changes at execution time. In this context the risk is amplified because the command may connect browser automation to a DevTools endpoint, so a malicious or trojaned package could interact with browser state, sessions, or local data.

Static analysis

No suspicious patterns detected.