bkn-requirement
Pass
Audited by Gen Agent Trust Hub on Jul 21, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or data exfiltration attempts were detected. The skill is designed to process business documents and transcripts to generate structured markdown outputs.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources such as meeting transcripts and customer materials.
- Ingestion points: Materials are archived in
projects/prj-<project>/inputs/round-XX/as defined inSKILL.md. - Boundary markers: The skill relies on structured project manifests and quality gate checklists (e.g.,
anti-drift-checklist.md) to maintain document integrity. - Capability inventory: The skill's operations are limited to file system read/write actions for documentation purposes. No capabilities for shell command execution, network requests, or privilege escalation were identified.
- Sanitization: The skill implements a multi-stage validation loop (Generate, Verify, Revise, Final Gate) to ensure that output remains consistent with provided business evidence.
Audit Metadata