skills/sztadii/agent-skills/full-pr/Gen Agent Trust Hub

full-pr

Pass

Audited by Gen Agent Trust Hub on Nov 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and follow instructions from the target repository, including contribution documents, pull request templates, and commit history. Maliciously crafted content in these files could attempt to override agent instructions or influence behavior during the automated PR process.
  • Ingestion points: Repository instructions, contribution docs, PR templates, and recent commit/PR history (SKILL.md).
  • Boundary markers: None explicitly defined to isolate repository content from system instructions.
  • Capability inventory: Git operations (commit, push), PR creation, and execution of arbitrary project build/start scripts.
  • Sanitization: No explicit sanitization or validation of the gathered repository instructions is specified.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to 'prepare the app' using 'the project's documented setup' (e.g., starting dev servers or building artifacts). This execution path presents a risk if the repository contains malicious build configurations or scripts that are executed in the host environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 7, 2026, 12:09 AM