The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external, potentially untrusted data such as codebases, architecture documents, READMEs, and ADRs.
Ingestion points: SKILL.md Step 1 and Step 2 involve reading codebase structures, deployment configurations, and requirement documents.
Boundary markers: There are no explicit delimiters or instructions provided in the templates to differentiate between agent instructions and potentially malicious content within the analyzed files.
Capability inventory: The skill has the capability to read local files, execute CLI analysis tools, and generate markdown reports.
Sanitization: While the skill includes a guardrail instruction to sanitize data before transmission to broad LLM models, it does not provide automated sanitization or validation logic for the input artifacts.
[COMMAND_EXECUTION]: The skill recommends the use of several external CLI tools for architecture analysis and fitness function verification.
Evidence: Instructions in SKILL.md and ARCHITECTURE_REVIEW_TEMPLATE.md suggest running madge --circular src/, npm audit, and dependency-cruiser to gather structural evidence.

arch-review