The Agent Skills Directory

[SAFE]: The skill is designed for text expansion and document generation, converting high-level specifications into detailed blueprints. It does not contain any instructions for network access, privilege escalation, or persistence.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided SPEC.md files to generate content. While this is an ingestion surface for untrusted data, the skill includes instructions to generate defensive implementation details, such as sanitization rules for potentially malicious paths like ../../etc/passwd.
[COMMAND_EXECUTION]: References to system operations (e.g., service startup, workspace management) are purely for documentation purposes within the generated blueprint and do not authorize the agent to execute shell commands.

blueprint-creator