The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): No arbitrary command execution patterns found. The validation script validate.py performs standard local filesystem checks and uses yaml.safe_load() for parsing.
[DATA_EXFILTRATION] (SAFE): No network operations or sensitive data access patterns were identified. The skill operates entirely on provided input data.
[EXTERNAL_DOWNLOADS] (SAFE): No external downloads, package installations (pip/npm), or remote scripts are referenced in the skill files.
[PROMPT_INJECTION] (SAFE): No instructions attempting to bypass safety filters, extract system prompts, or override agent behavior were found in the markdown or metadata.
[INDIRECT_PROMPT_INJECTION] (LOW):
Ingestion points: Processes requirements and current_architecture strings provided by the user.
Boundary markers: None explicitly defined in the schema.
Capability inventory: The skill is atomic and primarily used for recommendation and documentation; it lacks high-risk capabilities like network access or persistent file writes.
Sanitization: Standard schema validation is defined for structured inputs.

architecture-patterns