task-framing

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute a shell command using pbcopy to copy the final 'Session Brief' to the system clipboard. This is a local utility used for user convenience and does not involve network requests or access to sensitive system files.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from a local context file to inform the agent's questioning strategy.
  • Ingestion points: The agent is directed to read references/rami-context.md to understand the user's pillars and principles.
  • Boundary markers: The skill does not employ specific delimiters or 'ignore' instructions when reading the reference file.
  • Capability inventory: The skill has access to local shell command execution via pbcopy.
  • Sanitization: No validation or escaping is performed on the content of the reference file.
  • Note: The reference file is a static, local asset provided as part of the skill's own package, serving as an intended configuration source rather than untrusted external data.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 10:16 PM
Security Audit — agent-trust-hub — task-framing