write-abstract

Pass

Audited by Gen Agent Trust Hub on Jul 7, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted content such as user-provided abstract drafts and external venue metadata. This creates an indirect prompt injection surface where a draft could contain instructions intended to manipulate the agent's behavior. The risk is mitigated by the skill's reliance on deterministic scripts for validation and formatting.
  • Ingestion points: Abstract draft input described in SKILL.md and metadata fields processed by scripts/abstract_registration.py.
  • Boundary markers: No explicit boundary markers or 'ignore' instructions are used when processing draft content.
  • Capability inventory: File read/write access via Python scripts; execution of local Python utilities.
  • Sanitization: No specific input sanitization is implemented to prevent prompt injection in the ingested text.
  • [COMMAND_EXECUTION]: The skill invokes local Python scripts (abstract_check.py, abstract_registration.py, keywords_block.py) to perform its primary tasks. These scripts are self-contained, use only standard libraries, and do not execute arbitrary shell commands or require administrative privileges.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 7, 2026, 01:03 PM
Security Audit — agent-trust-hub — write-abstract