workflow-orchestration
Pass
Audited by Gen Agent Trust Hub on Mar 6, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill does not contain any executable scripts (such as Python, JavaScript, or Shell). It is composed entirely of Markdown files providing instructions and documentation for agent behavior.
- [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: The orchestration workflow documented in
WORKFLOW.md(Step 2: Context Discovery) involves reading project-level files such asCLAUDE.md,.specify/,plan/, andspecs/to gather context. - Boundary markers: The skill does not define specific delimiters to separate codebase context from its core instructions.
- Capability inventory: The skill references internal commands like
/core:developand/thinkwhich orchestrate multi-agent task execution and deep reasoning. - Sanitization: No specific sanitization or validation logic for the content read from external project files is provided in the skill instructions.
- Conclusion: This finding identifies a standard architectural surface for context-aware orchestration tools; no evidence of malicious behavior or exploitation was found.
Audit Metadata