Skills
Navigate
skills/deevus/skills/ste-plain-writing/Gen Agent Trust Hub

ste-plain-writing

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied technical documentation and prose which can contain untrusted instructions.
  • Ingestion points: Technical prose drafts provided via agent instructions or file reads (SKILL.md).
  • Boundary markers: The instructions lack explicit markers to isolate untrusted user content from agent logic.
  • Capability inventory: The agent is instructed to run a local Python script (scripts/ste_lint.py) to audit the text.
  • Sanitization: No specific sanitization or escaping of the input prose is performed before analysis.
  • [DYNAMIC_EXECUTION]: The test suite dynamically loads the linter script during execution.
  • Evidence: tests/test_ste_lint.py uses importlib.util to execute scripts/ste_lint.py relative to the test path.
  • [EXTERNAL_DOWNLOADS]: The skill documentation suggests cloning source code from GitHub for installation.
  • Evidence: README.md references the repository https://github.com/Ryuketsukami/ste-plain-writing.git.
  • [COMMAND_EXECUTION]: The skill requires the agent to execute a bundled Python utility to perform its primary function.
  • Evidence: SKILL.md contains instructions to execute python3 scripts/ste_lint.py path/to/draft.md.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 02:21 AM