sus-task
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a methodical process for task decomposition using a markdown-based template system. It does not exhibit malicious patterns such as remote code execution, unauthorized data access, or obfuscation.- [SAFE]: File system operations are appropriately scoped to a local artifact directory (~/.agents/artifacts/). The cleanup procedure ('Close') is a legitimate workflow step for disposing of transient workspace data and includes path resolution logic to prevent directory traversal.- [INDIRECT_PROMPT_INJECTION]: The skill identifies a potential attack surface by ingesting external specification files ('Suspec' artifacts). However, the risk is minimal as the instructions focus on structural decomposition rather than automated execution of untrusted commands.
- Ingestion points: Reads external specifications and source files mentioned in SKILL.md.
- Boundary markers: The skill uses structured templates (references/task-packet.md) to isolate requirement text.
- Capability inventory: Limited to file writing and deletion within the defined workspace directory.
- Sanitization: Workspace paths are resolved relative to the user's agent directory based on project identifiers.
Audit Metadata