project-task-flow

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides a structured framework for coordinating specialized agents. It includes explicit security and quality assurance steps, such as mandatory integration reviews and optional security audits for high-risk tasks.- [PROMPT_INJECTION]: The instructions define functional roles and operational logic for orchestration. There are no attempts to bypass safety filters, extract system prompts, or override agent constraints.- [DATA_EXFILTRATION]: No patterns of unauthorized data access or external transmission were found. The skill uses an internal progress ledger and local memory tools to manage workflow state without exposing sensitive credentials.- [REMOTE_CODE_EXECUTION]: The flow facilitates standard software development lifecycle commands (e.g., building and testing) within a multi-agent verification loop. There is no evidence of downloading or executing untrusted remote code.- [INDIRECT_PROMPT_INJECTION]: The skill processes task requirements ('the ask') through a multi-stage validation process involving a designer, implementer, verifier, and integration reviewer. This architecture significantly mitigates risks associated with untrusted data ingestion.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 07:50 AM
Security Audit — agent-trust-hub — project-task-flow