yzr-llm-workspace-management
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a local CLI tool,
llmw check-fixtures, to perform structural validation of the workspace. This tool is part of the vendor's workspace management ecosystem and is used for detecting configuration drifts. - [DATA_EXPOSURE]: The agent is authorized to read workspace metadata and wiki content. It contains specific instructions to avoid sensitive configuration files (such as those containing API keys) maintained by the CLI tool, thereby minimizing the risk of credential exposure.
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from multiple local wikis to generate summaries and answer queries. This ingestion of external data creates a potential surface for indirect prompt injection.
- Ingestion points: The skill reads
workspace.toml,wiki_metadata.toml, and all markdown content within thewiki/andraw/directories of registered wikis. - Boundary markers: The workflow relies on the
AGENTS.mdfile as a 'Single Source of Truth' and follows the structure defined inworkspace-spec.mdto scope its operations. - Capability inventory: The agent has the ability to write various workspace-level markdown files and execute the
llmwcommand-line utility. - Sanitization: There is no evidence of explicit content sanitization or filtering for the ingested wiki data before it is processed by the LLM.
Audit Metadata