ai-act-triage
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted content by scanning repository files using
grepto identify AI use cases. Maliciously crafted comments or metadata within the code could influence the agent's classification logic. - [INDIRECT_PROMPT_INJECTION]: The skill implements an 'overlay' mechanism that instructs the agent to check for and read external instruction files from a directory specified by the
POWERSUIT_OVERLAYenvironment variable. This allows external data to influence the agent's behavior and triage criteria at runtime. - [COMMAND_EXECUTION]: The skill utilizes shell commands (
grep,mkdir,cp,ls) to analyze the codebase and manage thecompliance/ai-register.mdfile. These operations are used for their intended purpose of inventory management and repository analysis. - [EXTERNAL_DOWNLOADS]: The skill references official legal documentation from
eur-lex.europa.eu. This is a well-known official government service and the reference is for informational purposes only.
Audit Metadata