multi-step-workflow-reporter
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill uses platform-provided orchestration and communication tools (define_subagent, invoke_subagent, send_message) as intended for workflow management and telemetry.
- [INDIRECT_PROMPT_INJECTION]: The skill processes telemetry data from subagents to update a live dashboard. While it lacks explicit output sanitization, the risk is negligible as the subagents are defined with fixed logic within the skill's own instructions.
- Ingestion points: Incoming telemetry payloads via send_message events in SKILL.md.
- Boundary markers: Absent.
- Capability inventory: Orchestration via define_subagent and invoke_subagent tools described in SKILL.md.
- Sanitization: Standard JSON parsing is performed without additional content filtering.
Audit Metadata