research-deep
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a Python script (
validate_json.py) located at/home/weizhena/.codex/skills/research/to verify research output. This is a vendor-provided tool used within the researcher's local environment. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests content from
outline.yamlandfields.yamlto generate prompts and command arguments without sanitization or boundary markers. - Ingestion points: Data is read from
outline.yaml(Step 1) andfields.yaml(Step 3) in the current working directory. - Boundary markers: No delimiters or safety instructions are used to wrap the interpolated YAML content in the prompt template.
- Capability inventory: The skill possesses the capability to execute shell commands via the
pythoninterpreter to run the validation script. - Sanitization: No validation, escaping, or filtering is applied to the YAML values before they are used to build the command-line string or the agent prompt.
Audit Metadata