apastra-validate

Pass

Audited by Gen Agent Trust Hub on Mar 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell scripts located within the user's local project directory to perform validation (e.g., bash promptops/validators/validate-prompt-spec.sh).
  • [PROMPT_INJECTION]: The skill processes untrusted data from various files within the promptops/ directory, which could potentially contain malicious instructions intended to influence the agent's behavior during the validation report generation.
  • Ingestion points: Files located in promptops/prompts/, promptops/datasets/, promptops/evaluators/, promptops/suites/, promptops/evals/, and promptops/policies/.
  • Boundary markers: None identified in the instructions.
  • Capability inventory: File reading, shell script execution via bash, and generation of validation reports.
  • Sanitization: The skill performs structural validation against JSON schemas but does not explicitly sanitize the content of the files to prevent the agent from interpreting embedded instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 15, 2026, 02:56 AM