code-review

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from story files and repository source code (references/workflow.md, Step 1 and 3) without explicit boundary markers or sanitization.
  • Ingestion points: Story files (Markdown) and application source code are read directly into the agent's context.
  • Boundary markers: Absent; the instructions do not define delimiters or warnings to ignore instructions embedded within the reviewed content.
  • Capability inventory: The agent is empowered to write to files to "Fix all HIGH and MEDIUM issues" and update project status files (references/workflow.md, Step 4 and 5).
  • Sanitization: Absent; no validation or filtering is applied to the ingested external content.
  • [COMMAND_EXECUTION]: The skill utilizes local git commands (git status, git diff) to discover changes in the project repository. These commands are used for metadata discovery and are consistent with the skill's primary function.
  • [SAFE]: The skill follows security best practices by explicitly excluding IDE and CLI configuration folders (e.g., .cursor/, .windsurf/, .claude/) from its analysis and modification scope, preventing accidental tampering with environment settings.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 09:05 PM