us-ac-formatter

Pass

Audited by Gen Agent Trust Hub on Sep 30, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied tables containing user stories and acceptance criteria (untrusted data) and reformats them into persistent documentation files.
  • Ingestion points: User-provided ASCII or markdown tables pasted into the agent's context.
  • Boundary markers: The instructions lack explicit delimiting or warnings to ignore instructions embedded within the user-supplied text.
  • Capability inventory: The skill writes multiple files to the docs/business/ directory and executes a local Python script.
  • Sanitization: No sanitization is performed on the input content; the skill is instructed to perform a "faithful conversion" and "preserve source language and exact domain terms."
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute a bundled Python script (scripts/build_ac_index.py) to maintain a cross-sprint index of acceptance criteria.
  • The script is located within the skill's own directory structure.
  • A review of the script's logic confirms it only uses standard Python libraries (os, re, glob, argparse) to read existing documentation files and write a summary index file.
  • No network calls, system-level modifications, or privilege escalation patterns were identified in the script's code.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 30, 2026, 02:49 AM