agent-global-context-capture

Pass

Audited by Gen Agent Trust Hub on May 11, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill directs the agent to interact with the local filesystem, including reading config.yaml and writing to ~/.agent-global-context/.
  • [PROMPT_INJECTION]: The skill manages untrusted conversation data, creating a surface for indirect prompt injection.
  • Ingestion points: Conversation excerpts and behavior descriptions (SKILL.md).
  • Boundary markers: Structured markdown template with specific metadata headers (SKILL.md).
  • Capability inventory: Local file read and write operations (SKILL.md).
  • Sanitization: Instructions to filter sensitive data and a prohibition against capturing secrets or private keys (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
May 11, 2026, 12:34 PM