Skills
skills/malinskibeniamin/skills/commit-push-pr/Gen Agent Trust Hub

commit-push-pr

Pass

Audited by Gen Agent Trust Hub on May 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes standard system binaries git and gh to manage repository state and interact with GitHub. It additionally calls a local script scripts/skills-browser.sh to facilitate screenshot capture for pull request documentation.\n- [DATA_EXFILTRATION]: Network communication occurs via git push and gh pr create to synchronize local code and metadata with remote repository hosting services, aligning with the primary functionality of the skill.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it analyzes untrusted data from the codebase (such as diffs and logs) to generate commit messages and PR descriptions.\n
  • Ingestion points: Repository data is ingested via git status, git diff, and git log commands in SKILL.md.\n
  • Boundary markers: Mitigation is provided through manual confirmation steps in Phase 1 and Phase 7, requiring the user to verify the scope and summary before completion.\n
  • Capability inventory: The agent possesses the capability to execute shell commands (git, gh) and perform network requests to the configured git remote.\n
  • Sanitization: No explicit sanitization or filtering is applied to the codebase content before it is interpolated into summaries or PR body templates.
Audit Metadata
Risk Level
SAFE
Analyzed
May 15, 2026, 09:07 AM
Security Audit — agent-trust-hub — commit-push-pr