skills/plurigrid/asi/cheapskate/Gen Agent Trust Hub

cheapskate

Pass

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [SAFE]: Comprehensive analysis of the 'cheapskate' skill confirms that its operations are confined to local text processing and the enforcement of efficiency guidelines. There are no network requests, sensitive file accesses, or obfuscated payloads present.
  • [COMMAND_EXECUTION]: The script scripts/evict.sh performs local text analysis using standard system utilities, including grep, wc, bc, and python3. These operations are limited to pattern matching and calculating text metrics (work ratio and compression ratio) for the provided input. No untrusted data is executed.
  • [PROMPT_INJECTION]: The skill provides instructions to the agent to adopt a direct and terse communication style, specifically advising the 'eviction' of preambles, apologies, and filler. These instructions are functional directives aligned with the skill's stated goal of token conservation and do not attempt to override safety guardrails or extract system prompt information.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 25, 2026, 07:38 PM