model-pr-diff-dossier

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is designed to ingest and analyze external content from GitHub pull requests, which presents a surface for indirect prompt injection.
  • Ingestion points: SKILL.md (Workflow steps 1 and 2) instructs the agent to collect PR links and read diffs directly from GitHub.
  • Boundary markers: The instructions do not provide delimiters or "ignore embedded instructions" warnings to protect the agent from malicious instructions that could be hidden within a PR description or code diff.
  • Capability inventory: The skill utilizes gh pr diff for data retrieval and mandates the execution of repository tests and formatting tools (SKILL.md step 7).
  • Sanitization: There are no requirements for sanitizing or validating the content retrieved from external PRs.
  • [COMMAND_EXECUTION]: The skill requires the agent to execute shell commands which could lead to unauthorized code execution if the local repository or the PR content being tested is malicious.
  • Evidence: SKILL.md (Workflow step 7) states "Run repository tests and formatting before publishing," which implies the execution of build or test scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 09:01 PM
Security Audit — agent-trust-hub — model-pr-diff-dossier