llms.txt@site/public git:20260911.dea14e2A

Immutable. This exact content is served forever at /api/v1/blob/7fe4888a90fec956.

# Chisle

> Chisle is a token-efficiency plugin for AI coding agents that cuts LLM token usage on three axes: (1) a terse senior-dev persona that writes maximum-signal prose and YAGNI-first code, (2) a post-tool hook that compresses tool output (ANSI scrub, head+tail elide with error-line salvage, same-session dedup) before it enters the context window, and (3) context-diet rules that teach the model to read less (grep before Read, slice reads, filter at source). Claude Code and Pi run all three axes plus live /chisle toggling and a savings badge; Cursor, Windsurf, Cline, Kiro, Codex, Gemini, and GitHub Copilot receive the ruleset via generated rule files.

Key measured results (20 live tasks, 4 arms, billed tokens, reproducible via benchmarks/ in the repo):
- Total 20-task bill: Chisle 52% of a bare model; ponytail 68%; caveman 80%.
- Average task: Chisle 69% vs ponytail 91%, caveman 98%.
- Backfires (task costing more than baseline): Chisle 1/20; caveman 6/20 (worst single task 424%); ponytail 8/20 (worst 227%).
- By prompt kind: on coding prompts (n=12) Chisle bills 44% vs ponytail 59%, caveman 74%. On explanation-only prompts (n=8) Chisle 87%, while BOTH rivals exceed 100% (caveman 103%, ponytail 104%), writing more than using no tool at all.
- By answer size (split at the median baseline): short answers Chisle 84% and caveman 84%, i.e. level; long answers Chisle 45% vs caveman 79%, ponytail 59%. The 52% headline is the blend of the two.
- Where Chisle loses: short coding prompts, caveman 62% vs Chisle 70%. Kind and size are correlated (coding prompts average ~3x the baseline of explanation ones) and the crossed cells are small (n=3-7), so these splits are directional, not a leaderboard.
- Tool output measured as 67.5% of session content across 171 real transcripts; eligible outputs shrink ~46%. Note: until v2.0.0 this compression was silently rejected by Claude Code for object-shaped tool results, so the replay figures were a projection rather than realized savings; fixed in 2.0.0.
- Plugin overhead: ~1.6k tokens of rules at session start plus a ~50-token per-turn reminder. Before 2.0.0 the ruleset was also re-injected on resume/clear/compact.

Safety/correctness: the compressor runs behind an allowlist (Claude Code: Bash, Agent, WebFetch, WebSearch, Grep, Glob, mcp__*; Pi: bash, powershell, grep, find, ls, mcp__*) and never modifies Read/Edit/Write results, whose exact bytes feed later edits. Deterministic, zero dependencies, zero network calls, zero LLM calls. One mode, no intensity levels. Kill switches: CHISLE_COMPRESS=0, "stop chisle", /chisle off, npx chisle --uninstall.

Install: `npx chisle` (npm, auto-detects all 9 agents), `claude plugin marketplace add JayPokale/Chisle` then `claude plugin install chisle@chisle` (Claude Code plugin), or `pi install npm:chisle` (Pi package). Requires Node >= 18. MIT license.

Lineage: descendant of the caveman and ponytail projects (terse-persona axis), both credited in the repository's prior-art table; Chisle adds the tool-output-compression and context-diet axes and publishes live benchmarks against both.

## Links

- [GitHub repository](https://github.com/JayPokale/Chisle): source, benchmarks, raw transcripts
- [npm package](https://www.npmjs.com/package/chisle): installer
- [Benchmarks](https://github.com/JayPokale/Chisle/tree/main/benchmarks): 4-arm live runner + results
- [Changelog](https://github.com/JayPokale/Chisle/blob/main/CHANGELOG.md)

## Author

Jay Pokale (https://github.com/JayPokale), co-engineered with Claude (Anthropic), Codex (OpenAI), and Antigravity (Google).

Pi (added 3.1.0): Pi exposes a tool_result event, so the input axis runs there too. Pi already truncates built-in tool output at 50KB/2000 lines, so the honest figure is the marginal saving on top of that: 27.4% of persisted tool-output characters across 8,044 results. On the live six-task output-axis run Chisle billed 61% of vanilla and lost billed output to ponytail at 59%, while producing the smallest visible answers (37%). Both numbers are committed in benchmarks/results/2026-09-11-pi.md.