blog-writer
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill displays an indirect prompt injection surface as it is designed to ingest and process research materials, links, and notes provided by the user.
- Ingestion points: Workflow Phase 1 in
SKILL.md(Requesting research materials, links, or notes). - Boundary markers: Absent. The instructions do not specify delimiters or instruct the agent to treat external content as untrusted.
- Capability inventory: The skill is capable of performing network writes via the Notion API (Phase 4) and local file operations, including writing to
references/blog-examples/and deleting files viamanage_examples.py. - Sanitization: Absent. No validation or filtering of ingested external content is mentioned.
- [COMMAND_EXECUTION]: The skill provides a Python utility,
manage_examples.py, which is used for pruning the examples library. This script possesses the capability to delete Markdown files (filepath.unlink()) within thereferences/blog-examples/directory. While used for library maintenance, this represents a functional local file system modification capability.
Audit Metadata