writing-shape

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted markdown content ('the pile') which is then used to generate article drafts. 1. Ingestion points: User-provided markdown input files as defined in SKILL.md. 2. Boundary markers: Absent; there are no specific instructions for the agent to ignore or delimit potentially malicious instructions within the raw input. 3. Capability inventory: File read and write operations. 4. Sanitization: Absent; the agent is instructed to rework fragments directly into the final article prose.
  • [SAFE]: The skill logic is focused on document preparation and formatting (e.g., lists, tables, callouts) and includes safeguards like re-reading files before writing to preserve user edits. No evidence of obfuscation, remote code execution, or unauthorized network activity was detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 10:17 AM
Security Audit — agent-trust-hub — writing-shape