hrt-align-consistency-review
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from multiple project artifacts and the codebase to perform its analysis. This creates a risk where malicious instructions embedded in these files could influence the agent's logic or the recommendations it provides to the user. * Ingestion points: discovery.md, proposal.md, specs/**/*.md, design.md, tasks.md, and general codebase content via search tools. * Boundary markers: The skill does not implement specific delimiters or 'ignore' instructions for the content it reviews, allowing for potential obedience to embedded commands. * Capability inventory: The skill executes shell commands (openspec, ripgrep, codegraph), invokes other skills, and spawns subagents for verification passes. * Sanitization: There is no evidence of semantic sanitization or instruction filtering for the content read from artifacts.
- [COMMAND_EXECUTION]: The skill invokes several external tools to perform documentation validation and codebase searches. These tools, including openspec, ripgrep, and codegraph, are used appropriately within the scope of the skill's primary function of technical documentation review.
Audit Metadata