logo-designer

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). The prompt includes explicit instructions to run subagents with mode: "bypassPermissions" (and similar directives to override platform controls), which are deceptive/unauthorized operational instructions outside the logo-design purpose.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill explicitly instructs the agent to read and extract content from user-provided repositories or URLs (e.g., "Read the README, package.json, CSS/config files" in "Step 1: Gather context automatically") and to clone/inspect repos in "Phase 5: Repo Integration", which are untrusted third-party sources whose contents the agent is expected to interpret and use to drive design decisions and tooling actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The export script invokes npx (e.g., "npx --yes @aspect-build/resvg" and suggests "npm install -g @aspect-build/resvg"), which will fetch and execute code from the npm registry at runtime to perform SVG→PNG conversion, so there is a runtime external dependency that executes remote code.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The skill explicitly instructs running local scripts, cloning/modifying repositories and committing files and—most notably—spawning subagents with mode="bypassPermissions", which directs the agent to bypass permissions and modify the machine's state.