review-pr

Pass

Audited by Gen Agent Trust Hub on Nov 19, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources. Evidence Chain: 1. Ingestion points: pull request content is read from pr_description.txt and pr_diff.txt. 2. Boundary markers: The instructions do not define explicit delimiters or instructions to ignore embedded commands within the ingested text. 3. Capability inventory: The skill writes to review.json and executes the scripts/validate_review_json.py script. 4. Sanitization: No sanitization of the input text is specified.
  • [COMMAND_EXECUTION]: The workflow requires executing a local script scripts/validate_review_json.py to ensure the output matches the required schema. Analysis of the script indicates it uses only standard Python libraries (json, re, sys, pathlib) and performs structural validation without network access or sensitive file interaction.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 19, 2026, 01:13 AM