context-curation

Pass

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • Ind irect Prompt In j e c t i o n (LOW): The skill is vulnerable to ind irect prompt in j e c t i o n because it ingests untruste d data from git changes and proj e c t
  • s p e c i f i c i nstruction file s with out sanitizatio n.
  • Ingestio n point s: Read s git dif f out put (Step 1 & 2) and various local context file s lik e .cursor rule s, CLAUDE.m d, or .aider.conf.ym l (Step 4).
  • Boundary marker s: None presen t ; the skill doe s not use delimiter s or 'ignore' prompt s to isolate untruste d conten t .
  • Capabilit y invent or y: The skill onl y perform s rea d
  • onl y file system and rea d
  • onl y git operation s.
  • Sanitizatio n: Absen t ; the skill doe s not filt e r or escape conten t from the code base before processin g it.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 19, 2026, 03:03 PM