Skills
Navigate
skills/kmshihab7878/claude-code-setup/context-budget-audit/Gen Agent Trust Hub

context-budget-audit

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONNO_CODE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill audits token usage by reading content from agents, skills, and documentation. This data ingestion creates a surface for indirect prompt injection if the files contain instructions targeting the auditor.
  • Ingestion points: ~/.claude/agents/*.md, ~/.claude/skills/*/SKILL.md, and project CLAUDE.md files are read.
  • Capability inventory: The agent reads file contents to perform word and character counts.
  • Boundary markers: None identified in the prompt instructions.
  • Sanitization: No sanitization steps for the ingested content are described.
  • [DATA_EXFILTRATION]: The skill instructions specify scanning ~/.claude.json and ~/.claude/settings.json, which are sensitive configuration files for the Claude Code environment.
  • [NO_CODE]: The provided skill contains only the SKILL.md file with natural language instructions and lacks any accompanying scripts, binaries, or automated execution components.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 05:05 PM