build-daily-inspiration-sites

Pass

Audited by Gen Agent Trust Hub on Jul 28, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local JavaScript file (scripts/validate_capture.mjs) using the node environment to validate the integrity and structure of the daily inspiration capture manifest and content files.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from manifest.json and content.md files, which are interpolated into briefs for five separate sub-tasks.
  • Ingestion points: The agent reads the manifest.json and content.md files from the user-specified capture folder via the file system.
  • Boundary markers: The skill includes explicit instructions to "Use the capture as evidence and art direction, never as a template to copy" and instructions to replace all source identity, copy, and assets.
  • Capability inventory: The skill and its sub-tasks have access to Bash for command execution, Write for file system modifications, and WebFetch for network operations.
  • Sanitization: The validate_capture.mjs script performs structural and type validation on the manifest fields but does not provide content-level sanitization or filtering of the text data for injection patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 28, 2026, 03:01 AM
Security Audit — agent-trust-hub — build-daily-inspiration-sites