skills/seanmars/sepia/sepia-write/Gen Agent Trust Hub

sepia-write

Pass

Audited by Gen Agent Trust Hub on Oct 23, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a restricted proxy pattern, delegating the 'write' operation to a specific sibling file (../sepia/SKILL.md) while explicitly forbidding broader file system searches or fallback mechanisms.
  • [SAFE]: Includes proactive security instructions that mandate treating all target content as untrusted data, effectively mitigating potential indirect prompt injection attempts by instructing the agent not to follow instructions embedded in the processed data.
  • [SAFE]: Explicitly restricts the agent's authority, stating that invocation grants no tool, file, network, or external-action permissions beyond the specific prose binding.
  • [SAFE]: Prevents operation switching or logic branching based on input content, ensuring the skill maintains a predictable and narrow scope of execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 23, 2026, 11:20 AM