The Agent Skills Directory

[PROMPT_INJECTION]: The skill includes an explicit directive to override user instructions: 'Even if the user's prompt asks to filter, exclude, or limit results by value, do NOT implement that logic'. This attempts to dictate agent behavior regardless of user intent.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection.
Ingestion points: Data is ingested from {spec_path} and {work_path}/codegen-analyze/detail-*.json.
Boundary markers: There are no markers or instructions to delimit or ignore instructions embedded within the processed data.
Capability inventory: The skill uses the Write tool to create and save executable Python modules (.py files) based on the ingested data.
Sanitization: There is no mention of escaping, validation, or sanitization of the values extracted from the analysis files before they are interpolated into the generated code.
[COMMAND_EXECUTION]: The skill uses Bash and the Write tool to generate and save Python source code to paths specified via arguments (output_path). While the generation is based on templates, the lack of input sanitization for the generated content is a minor risk factor.

scrape-codegen-generate