create-pull-request

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses several standard Git commands to gather necessary information for the pull request. It executes git remote get-url origin to identify the hosting platform, git branch -a to list available branches, and git diff / git log to summarize the code changes being submitted.
  • [EXTERNAL_DOWNLOADS]: In Step 4, the skill is instructed to fetch content from user-provided ticket URLs. This involves making network requests to external domains to retrieve task descriptions and requirements for the PR summary.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because it processes untrusted data to generate its output.
  • Ingestion points: External content from ticket URLs and markdown planning files within the local workspace (SKILL.md, Step 4).
  • Boundary markers: The instructions do not define specific delimiters or "ignore previous instructions" markers when interpolating the fetched ticket or file content into the PR template.
  • Capability inventory: The agent can perform network requests, execute git commands, and utilize platform-specific MCP tools to create and submit pull requests (SKILL.md, Step 6).
  • Sanitization: There is no mention of sanitizing or escaping the retrieved external text before it is used to synthesize the PR description.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 02:55 PM
Security Audit — agent-trust-hub — create-pull-request