illustrate
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input through the $ARGUMENTS variable and passes it to downstream components like diagram-brief and drawing skills.
- Ingestion points: The $ARGUMENTS variable in SKILL.md (containing titles, notes, or brief paths).
- Boundary markers: No explicit delimiters or instructions are provided to the agent to treat the embedded data as untrusted.
- Capability inventory: The skill coordinates file writes (writing brief.md) and complex rendering logic across multiple sub-skills.
- Sanitization: There is no evidence of input validation or sanitization before passing user data to sub-agent tools.
- [COMMAND_EXECUTION]: Documentation for the skill includes examples of shell commands for headless execution via the acpx CLI. It suggests the use of the --approve-all flag, which overrides security guardrails that normally require human confirmation for file system operations or tool execution.
- [EXTERNAL_DOWNLOADS]: The skill documentation references Paldom/diagram-skills as an external dependency to be installed via the npx package runner, which is a vendor-provided resource necessary for the orchestration workflow.
Audit Metadata