proposal-writing

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.90). The skill's "Companion check" explicitly instructs the agent to run shell commands (ls of user skill dirs and an npx install) to probe the local filesystem and offer installs, which is an operational action outside the proposal-writing purpose and poses a prompt-injection/privacy risk.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill's companion check tells the agent at first activation to run "npx skills add AbsolutelySkilled/AbsolutelySkilled --skill " which will fetch and execute remote code via npx at runtime, creating a clear remote-execution dependency.