archetype-review-base
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The framework is designed to ingest and analyze external documentation (ARCH docs) and source code, creating a surface for indirect prompt injection. The skill mitigates this by mandating a quote-verification process using a utility script to ensure findings are grounded in actual source text.
- [DYNAMIC_EXECUTION]: The skill uses a shell-based discovery pattern to locate its own utility scripts within the platform's plugin cache (e.g.,
~/.claude/plugins/cache/). This involves executing a Node.js script from a path calculated at runtime via shell sub-expressions. - [COMMAND_EXECUTION]: The skill utilizes a project-specific task management tool (
bd) and custom shell scripts (log-verdict.sh) to manage the review lifecycle and log results. These operations are restricted to specific, defined tasks within the review workflow.
Audit Metadata