workspace
Pass
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill definition consists solely of a Markdown file (
SKILL.md) and does not include any executable code, scripts, or binary files. - [PROMPT_INJECTION]: The documentation describes a mechanism for reading files generated by other agent tools, which serves as a surface for indirect prompt injection where the agent could potentially process untrusted instructions embedded in the data.
- Ingestion points: The
workspace_readfunction as described inSKILL.mdallows the agent to ingest content from paths such ascode-agent/anddocuments/. - Boundary markers: No delimitation or 'ignore instructions' warnings for the ingested content are specified in the documentation.
- Capability inventory: The skill provides functions for listing, reading, and writing files within the shared session workspace.
- Sanitization: There is no documentation regarding the sanitization or validation of content read from the workspace.
Audit Metadata