verify-ui-change-in-cloud

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The skill directs spawning a cloud agent with computer-use enabled and explicitly instructs hardcoding or removing channel/feature gates (bypassing security checks) to exercise code paths for testing, which weakens security boundaries and is suspicious.

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly requires embedding the exact cloud environment_id (SVhg783GBFQHk1OfdPfFU9) in the run_agents invocation/config, which forces the LLM to output that secret-like identifier verbatim and could expose it.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs a cloud agent to clone the repository and build/run its code (via git clone implied and "cargo run") and explicitly checks the Git remote contains warpdotdev/warp, so at runtime it will fetch and execute code from the external repo warpdotdev/warp.

HIGH W008: Secret detected in skill content (API keys, tokens, passwords).

• Secret detected (high risk: 1.00). The document contains a high-entropy, literal environment identifier: SVhg783GBFQHk1OfdPfFU9. It is directly present (not a placeholder) and is used as the "environment ID" for the Warp Dev Environment (used in remote.environment_id), which appears to be a real identifier that could grant access or be used to spawn cloud agents. That meets the definition of a secret (high-entropy value that can provide access), so it is flagged.