writing-specs

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill mentions the use of @ara-commons/ara-skills via npx. This is documented as a prerequisite for a specific optional research packaging stage and is consistent with the skill's stated purpose of research automation. The source is a known namespace within the research automation domain.
  • [COMMAND_EXECUTION]: The instructions refer to npx @ara-commons/ara-skills. This is a documented dependency for compiling research results and is presented as a manual installation requirement rather than an automated hidden command.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user input (North Star, ResearchBrief) to generate a specification. It includes a mandatory spec-self-review step and multiple user approval gates (Step 3 and Step 6) to mitigate the risk of malicious instructions in the input influencing the final executable specification.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 12:04 AM