mobile-ci-cd
Pass
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill uses explicit directives to override standard agent behavior, such as 'You are in AUTONOMOUS MODE', 'Do NOT ask questions', and 'Do NOT pause for confirmation'. These instructions aim to bypass human-in-the-loop validation for the creation of security-sensitive CI/CD configurations.
- [PROMPT_INJECTION]: The skill ingests untrusted external data from project files (e.g., package.json, pubspec.yaml, build.gradle.kts) to drive its generation logic, creating a surface for indirect prompt injection.
- Ingestion points: Reads framework-specific configuration files during Phase 1 (SKILL.md).
- Boundary markers: Absent; the skill does not use delimiters or instructions to ignore embedded directives within the ingested files.
- Capability inventory: The skill is capable of writing multiple workflow files to the filesystem and executing shell-based validation commands (Phase 7).
- Sanitization: No sanitization or validation of the ingested content is specified before it is used to generate or validate configurations.
- [COMMAND_EXECUTION]: The skill performs shell-based validation of generated assets in the 'SELF-HEALING VALIDATION' phase, including commands like 'docker build --check' and 'terraform validate'. While these are used for correctness, they execute in the user's environment based on generated content.
Audit Metadata