ai-context-refresh

Pass

Audited by Gen Agent Trust Hub on Nov 10, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill generates the project context by reading docs/project-config.json and documentation files, which could serve as a vector for indirect prompt injection if these files contain malicious content.\n
  • Ingestion points: docs/project-config.json and markdown files within the docs/ directory.\n
  • Boundary markers: The skill uses markers to identify and update managed content sections.\n
  • Capability inventory: The skill performs filesystem read/write operations and Node.js script execution.\n
  • Sanitization: The renderCredentialReference function in section-builders.cjs redacts secrets like passwords and API keys from configuration data. Markdown escaping is used to prevent structure injection in tables.\n- [COMMAND_EXECUTION]: The skill workflow requires the execution of Node.js scripts (generate-claude-md.cjs and run-codex-sync.mjs) to automate context updates.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 10, 2026, 08:16 PM