The Agent Skills Directory

[PROMPT_INJECTION] (LOW): Indirect Prompt Injection (Category 8) vulnerability surface detected.
Ingestion points: The skill is designed to process user-provided Dockerfiles for review and refactoring tasks as specified in the trigger conditions in SKILL.md.
Boundary markers: There are no defined delimiters or instructions to ignore embedded prompts within the analyzed Dockerfile content.
Capability inventory: The agent has the capability to generate and modify complex Dockerfiles containing shell scripts and system configurations (SKILL.md).
Sanitization: No sanitization or validation of external Dockerfile content is performed before processing.
[EXTERNAL_DOWNLOADS] (LOW): Documentation includes examples of using curl and wget to download scripts and tools (e.g., in references/lint-pipefail.md and references/inst-copy-over-add.md). These examples use placeholder domains (example.com) or well-known third-party sources (nodesource.com) for educational purposes.
[COMMAND_EXECUTION] (LOW): Numerous examples of shell commands like apt-get, pip, and npm are present in the reference files. These are intended for use in Dockerfile construction and are not executed by the skill itself.
[CREDENTIALS_UNSAFE] (SAFE): Documentation includes dummy secret strings (e.g., 'sk-live-abc123' in references/sec-secret-mounts.md) to demonstrate how to properly manage sensitive data. No real credentials or private keys were found.

dockerfile-optimise