# Behavioral Impact

> Removed the unused ToolBuffer shim from the agent loop and strengthened the system prompt to demand batched read-only tool calls during discovery, aligning the runtime with actual parallel tool execution capabilities.

- Skill: `tools-only/behavioral-impact-8` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add tools-only/behavioral-impact-8`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tools-only/behavioral-impact-8/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: tools-only (https://skillmd.com/u/tools-only)
- Updated: 2026-09-29
- Page: https://skillmd.com/skills/tools-only/behavioral-impact-8

---


## Summary
Removed the unused ToolBuffer shim from the agent loop and strengthened the system prompt to demand batched read-only tool calls during discovery, aligning the runtime with actual parallel tool execution capabilities.

## Context
ToolBuffer had no call sites and was only flushed at request completion, so it never influenced execution. Prompt guidance needed explicit batching language to reduce sequential read loops.

## Changes
- Removed ToolBuffer file, exports, and call wiring from the agent loop and orchestrator signature.
- Updated parallel execution guidance in the system prompt with explicit batching requirements and an example.

## Behavioral Impact
- No runtime behavior change from ToolBuffer removal (buffer was unused).
- Model guidance now more forcefully requests batched tool calls after discovery.

## Related Cards
- [[2026-02-02-tool-name-normalization]]