agent-patterns
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists of markdown documentation and instructions for an agent-based workflow. It does not include executable scripts, network operations, or sensitive data access. External links in the reference files point to the author's official domains.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow where agents ingest external data, such as specification files and reports.
- Ingestion points: The agent reads specification files (.claude/sprint/...), API contracts, and structured agent reports (SKILL.md, references/examples.md).
- Boundary markers: The instructions mandate the use of SPAWN REQUEST and AGENT REPORT tags as delimiters to separate coordination instructions from implementation data (SKILL.md).
- Capability inventory: The skill is limited to the 'Read' tool and output generation for a separate sprint orchestrator; it does not perform file writes, network calls, or shell execution.
- Sanitization: The skill relies on structural formatting and domain partitioning to manage external content rather than explicit code-level sanitization.
Audit Metadata