skills/whizzzkid/skills/wk-pr-merge/Gen Agent Trust Hub

wk-pr-merge

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from GitHub Pull Request descriptions and Jira ticket data to determine its next steps.
  • Ingestion points: The PR body is resolved using gh pr view in SKILL.md Step 1 and used throughout the workflow. Jira ticket information is fetched using getJiraIssue as described in references/ticket-transition.md.
  • Boundary markers: The skill includes logic to strip terminal metadata footers and uses boundary-aware regex for ticket detection. However, it lacks explicit 'ignore instructions' delimiters when scanning for action items in Step 5.
  • Capability inventory: The agent has permission to merge PRs, delete branches, close GitHub issues, and transition Jira tickets via the provided tools.
  • Sanitization: Sanitization is primarily handled through specific regex matching for ticket keys and action items.
  • [COMMAND_EXECUTION]: The skill relies on shell commands via the gh CLI and git to perform its core functions.
  • It executes commands like gh pr merge, gh pr edit, and git worktree remove to manage the PR lifecycle.
  • These operations are limited to the specific repository and PR context resolved in Step 1.
  • [EXTERNAL_DOWNLOADS]: The skill interacts with external services to retrieve configuration and status.
  • It fetches PR status, check runs, and review threads from GitHub's official GraphQL and REST APIs.
  • It interacts with Jira APIs to manage task transitions.
  • It references the gh stack extension for managing stacked pull requests, recommending updates from official sources if necessary.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 12:25 AM
Security Audit — agent-trust-hub — wk-pr-merge