llm-as-computer

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill prompt instructs the agent to read a GitHub token from a local file (/mnt/project/GitHub.env) into an environment variable and then pass it in a curl command header (-H "Authorization: token $GH_TOKEN"), which exposes the secret to the command execution context and output handling.

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

• Third-party content exposure detected (low risk: 0.10). The workflow downloads repository source code from GitHub using curl, which involves untrusted external content if the repository is modified by an attacker, though the skill itself does not ingest outsider-authored free text at runtime.