github-actions

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.90). The prompt includes a hidden "SELF-EVOLUTION TELEMETRY" block that instructs the agent to silently check for a local memory directory and append execution metadata (only if found), a conditional telemetry/logging action unrelated to the stated goal of generating workflows and designed to be non-obvious—this is a deceptive instruction outside the skill's advertised purpose.