cmd-write-proofread

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted text and is authorized to modify files based on that content.
  • Ingestion points: The agent reads the "full post" which is external, untrusted data (found in SKILL.md instructions).
  • Boundary markers: There are no instructions for the agent to use delimiters or ignore embedded directives within the post being proofread.
  • Capability inventory: The agent is granted the ability to "edit the file directly" (SKILL.md, Instruction 5), providing a mechanism for an injection to result in unauthorized file system changes.
  • Sanitization: The skill lacks any input validation or sanitization routines for the content it processes.
  • [NO_CODE]: No scripts or executable files are included with this skill; it relies entirely on instructions to the AI agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 10:07 AM
Security Audit — agent-trust-hub — cmd-write-proofread