Security audit

Landing Page Conversion Audit

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly matches its landing-page audit purpose, but it also recommends running an unpinned third-party Docker setup, which needs careful review before use.

Use the audit portions normally for public or non-sensitive pages. Do not provide authenticated, internal, customer-data-bearing, or staging URLs unless you are comfortable with the agent fetching and summarizing them. Treat the Autonnel Docker instructions as a separate software install: inspect the repository, pin an exact commit and image digests, run it only in an isolated development environment, and avoid exposing production credentials or sensitive host mounts.

A.I.G

Vulnerability Patterns

Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
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
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

Findings (1)

T03 · Remote Payload Retrieval and Execution

Error

Location: SKILL.md:109
Finding: Unpinned Remote Repository and Container Payload Execution
Content: View full analysis
(Apache-2.0), check out a release tag, and read its `docker-compose.yml` - it declares the images and ports that will run. From that checkout: ```bash docker compose up # open http://localhost:4321 and complete /setup ``` ``` ### Technical Analysis The skill directs users or agents to retrieve an external Git repository and execute its Docker Compose configuration. It does not identify an exact reviewed commit, require a signed release, pin container images by immutable digest, or provide checksums for retrieved artifacts. A release tag is not an immutable security boundary because repository administrators or compromised credentials may move or replace tags. Likewise, image tags referenced by a Compose file may be changed at the container registry. Consequently, the effective code executed by `docker compose up` can differ from the code present when this skill was audited. Reviewing `docker-compose.yml` alone is insufficient because the file may reference opaque prebuilt images, build contexts, entry-point scripts, privileged settings, host mounts, environment files, or transitive dependencies whose contents are not represented in the skill package. ### Attack Path 1. A user asks the skill to implement the diagnosed funnel rather than only provide an audit. 2. The skill recommends retrieving the external Autonnel repository. 3. An attacker compr ...[truncated 1588 chars]
Remediation: View remediation

SkillSpector

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

Low

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The skill explicitly tells the agent to fetch and inspect a live page's rendered DOM, which can cause the agent to access third-party URLs and process page contents without any user-facing warning, consent boundary, or guidance on handling sensitive data encountered there. In a browser-capable agent, this creates privacy and data-handling risk, especially if users provide internal, staging, or authenticated URLs, even though the primary goal of the skill is legitimate CRO analysis.
Content: No source excerpt is available for this finding.