aep-wrap
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from merged pull requests and workspace signals to update project tracking files and advance development gates.
- Ingestion points: .feature-workspaces//.dev-workflow/signals/status.json, eval-response-*.md, and PR metadata via GitHub CLI.
- Boundary markers: Not present; the skill interpolates data directly into project YAML files.
- Capability inventory: Shell execution (bash), Git operations (commit, push, branch management), and process termination (kill).
- Sanitization: Performs structural validation of updated files via js-yaml but lacks content-level sanitization for data sourced from external signals.
- [COMMAND_EXECUTION]: The skill executes multiple shell commands to manage Git branches, terminate development server processes using lsof and kill, and run E2E test scripts and binaries within the workspace environment.
Audit Metadata