dsh-merging-stacked-prs

Pass

Audited by Gen Agent Trust Hub on Sep 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) and the gh stack extension to automate the merging of dependent pull requests. The commands are scoped to standard repository management tasks.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external metadata from GitHub, representing a potential injection surface.
  • Ingestion points: PR metadata such as author names, branch names, and stack positions are fetched via gh pr view and the GitHub GraphQL API as described in SKILL.md.
  • Boundary markers: The instructions mandate strict verification of author.login across the stack and require the agent to use the official GitHub stack object as the authority for ordering.
  • Capability inventory: The skill utilizes gh stack link, gh stack sync, and gh stack merge for repository mutations.
  • Sanitization: It requires manual user confirmation if authors differ or stack metadata is inconsistent, and it enforces a hard-stop on cross-repository chains.
  • [SAFE]: The skill follows security best practices by avoiding force-pushes, requiring local validation after stack synchronization, and ensuring PRs are fully merged before deleting branches.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 10, 2026, 01:54 AM
Security Audit — agent-trust-hub — dsh-merging-stacked-prs