project-brain

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes git diff --name-only in assets/modes/SAVE.md to identify files modified during a session for logging purposes. This is an expected behavior for a project context tool.- [PROMPT_INJECTION]: Potential surface for Indirect Prompt Injection. The skill reads external markdown files (LOAD.md) and presents a summarized briefing to the agent. Malicious content in these files could theoretically influence agent behavior. * Ingestion points: assets/modes/LOAD.md Step 3 reads documents from the local filesystem or Obsidian vaults. * Boundary markers: Absent. Content is parsed for headings but not delimited with security warnings. * Capability inventory: Read, Write, Edit, Glob, Grep, ToolSearch, AskUserQuestion. * Sanitization: None. The skill does not filter natural language content from loaded files.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 07:31 AM