promote

Pass

Audited by Gen Agent Trust Hub on Mar 23, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill creates an indirect prompt injection surface by reading data from multiple project-specific memory files and promoting that content to the global agent configuration. 1. Ingestion points: Extracts information from ~/.claude/projects/*/memory/*.md and project-level CLAUDE.md files. 2. Boundary markers: Absent; no specific delimiters or ignore-instructions are used when content is interpolated into the global config. 3. Capability inventory: The skill possesses file-read and file-write capabilities targeting the global ~/.claude/CLAUDE.md file, which influences agent behavior across all sessions. 4. Sanitization: Absent; the skill relies on semantic comparison but does not filter or sanitize the promoted content.
  • [COMMAND_EXECUTION]: The skill performs automated local file system discovery and directory traversal. Evidence: It lists all subdirectories within ~/.claude/projects/ and identifies specific markdown files for reading and aggregation.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 23, 2026, 10:58 AM
Security Audit — agent-trust-hub — promote