skills/deevus/tea-skills/create-pull/Gen Agent Trust Hub

create-pull

Pass

Audited by Gen Agent Trust Hub on Jun 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill facilitates pull request management by executing the tea CLI and git command-line tools. It also invokes a local Python script actions/pull-requests/find-by-branch.py to automate pull request discovery.- [PROMPT_INJECTION]: The skill exposes an indirect prompt injection surface via the ingestion of untrusted commit metadata.
  • Ingestion points: Commit messages are retrieved using git log -1 --format=%s and used to populate pull request titles in the 'Quick Pattern' example.
  • Boundary markers: None identified; the commit message is interpolated directly into the shell command argument for --title.
  • Capability inventory: The agent has the capability to execute shell commands (tea, git) and run local Python scripts.
  • Sanitization: There is no evidence of sanitization or escaping of the commit message before its use in a shell command, which could potentially lead to command injection if a malicious commit message contains shell metacharacters.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 20, 2026, 09:18 AM
Security Audit — agent-trust-hub — create-pull