openclaw-design-audit

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses shell commands to manage repository state and interact with GitHub APIs.
  • Evidence: Uses git checkout, git fetch, git reset, and git push --force-with-lease to manage automation branches in references/github-pr-delivery.md.
  • Evidence: Uses gh pr create and gh pr edit to automate the delivery of audit findings to the openclaw/clawhub repository.
  • [EXTERNAL_DOWNLOADS]: The skill performs installations of external components required for the audit workflow.
  • Evidence: The workflow in references/github-pr-delivery.md specifies the installation of 'Carapace' at a pinned Git tag.
  • [PROMPT_INJECTION]: The skill is designed to process untrusted source code, creating a surface for indirect prompt injection.
  • Ingestion points: Audits consumer-provided frontend code, commit SHAs, and rendered routes as defined in SKILL.md.
  • Boundary markers: The skill does not explicitly define delimiters to separate audited code from its own instructional context.
  • Capability inventory: The agent has the ability to write reports to the filesystem and push code changes/PRs to GitHub.
  • Sanitization: The instructions do not specify sanitization or escaping of the audited code before it is processed by the agent's reasoning engine.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 01:28 AM
Security Audit — agent-trust-hub — openclaw-design-audit