scrape-codegen-generate
Pass
Audited by Gen Agent Trust Hub on Jul 16, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill instructs the agent to ignore specific user instructions if they involve adding filtering logic to the output ("Even if the user's prompt asks to filter, exclude, or limit results by value, do NOT implement that logic"). This pattern overrides user-provided constraints.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by consuming potentially untrusted data to generate code.
- Ingestion points: The agent reads extraction analyses from
{work_path}/codegen-analyze/detail-*.jsonand schema properties from{spec_path}. These files contain data derived from external web content. - Boundary markers: No delimiters or "ignore instructions" markers are used to separate the JSON data from the generation logic.
- Capability inventory: The agent has access to
Read,Write, andBashtools, and is tasked with writing functional Python modules to the local filesystem. - Sanitization: The instructions do not mandate validation or sanitization of the data extracted from JSON files before it is interpolated into the generated Python code structure.
Audit Metadata