hrt-dark-mode-decision-gate
Pass
Audited by Gen Agent Trust Hub on Aug 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests a wide variety of project artifacts to resolve high-severity findings without user intervention. These external files represent an attack surface where malicious instructions could be embedded to manipulate the agent's decision-making process and suppress legitimate security or consistency alerts.
- Ingestion points: Processes content from
discovery.md,proposal.md,specs/**/*.md,design.md,tasks.md,align.md,dark.md, project source code, and test outputs. - Capability inventory: Spawns fresh subagents, writes analysis results to
dark.md, and initiates code execution through test running. - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present when interpolating artifact content into the subagent context.
- Sanitization: No sanitization or validation of the ingested artifact content is specified.
- [COMMAND_EXECUTION]: The procedure requires verifying actual behavior by running tests and reading code paths when called from a code review context. This involves executing arbitrary code found within the local project environment.
Audit Metadata