Chisle

Chisle is a Claude Code plugin that makes Claude cheaper to run without making it dumber. It cuts what Claude writes: no filler, no hedging, no speculative abstractions, just the smallest code that works. It also cuts what Claude reads: a PostToolUse hook trims oversized tool output by ~46% before it re-enters the context window, where it would be re-billed on every later request. In agent-loop tests Claude with Chisle passes the same tasks as Claude without it. One npx chisle installs it, and the same ruleset ships to Pi, Cursor, Codex, Gemini, Copilot, OpenCode, Antigravity and four more agents. Every number here comes from committed raw transcripts, including the runs where it lost.

---

"Implement an order state machine covering created, paid, shipped, delivered, cancelled and refunded, with the legal transitions enforced and invalid ones rejected." Same model, same prompt, one difference: the injected ruleset. Both answers are the verbatim committed output from raw-oct-notools/large-s2:

An Enum of states, a custom exception class, type hints and docstrings on every method, a file write, then a test harness:

class OrderState(Enum):
    CREATED = "created"
    PAID = "paid"
    # … four more

class InvalidTransition(Exception):
    """Raised when attempting an invalid state transition."""
    pass

class Order:
    VALID_TRANSITIONS = { … }
    def transition_to(self, next_state: OrderState) -> None:
        """Transition to next state if valid, …"""

…plus usage, tests and a summary.

One dict of legal transitions, one method that enforces it:

class OrderStateMachine:
    TRANSITIONS = {
        'created': {'paid', 'cancelled'},
        'paid': {'shipped', 'refunded'},
        'shipped': {'delivered', 'cancelled'},
        'delivered': {'refunded'},
        'cancelled': set(),
        'refunded': set(),
    }
    def transition(self, new_state):
        if new_state not in self.TRANSITIONS[self.state]:
            raise ValueError(f"Invalid transition: …")
        self.state = new_state

Then a short usage block and one line on how it works.

Same behaviour, a third of the tokens. caveman wrote 1,140 and ponytail 842 on the same prompt.

What it compresses

axis	what	how
Output: prose	filler, hedging, manufactured structure	zero-fluff ruleset, injected per session
Output: code	speculative abstractions, unrequested boilerplate	YAGNI efficiency ladder
Input: context	oversized tool output flooding the window	Claude PostToolUse, Pi tool_result, OpenCode plugin, Copilot CLI postToolUse: scrub, elide, dedup, plus prevention rules
Input: MCP	oversized MCP tool results in any agent	npx chisle mcp wrap: a proxy in front of each stdio MCP server (Cursor, Windsurf, Cline, Kiro, Gemini, Qwen Code, Copilot CLI)

Where each one attaches to a session:

flowchart LR
    subgraph S["Session start"]
        H1["Ruleset injection<br/>once per active session"]
    end
    subgraph T["Every turn"]
        H2["Mode tracking<br/>Claude + Pi"]
    end
    subgraph L["Every tool call"]
        H3["PostToolUse / tool_result<br/>scrub → elide → dedup"]
    end

    H1 --> M(["Model"])
    H2 --> M
    M -->|writes| O["Output:<br/>terser prose,<br/>YAGNI-first code"]
    M -->|calls a tool| TOOL[["Bash / grep / web / extension tools"]]
    TOOL -->|raw output| H3
    H3 -->|"compressed, rebuilt into<br/>the tool's own shape"| M

    RE["Read / Edit / Write"] -.->|"never touched,<br/>exact bytes feed later edits"| M

    style M fill:#1f2937,stroke:#d78a3c,color:#e6edf3
    style O fill:#14532d,stroke:#2da44e,color:#e6edf3
    style H3 fill:#1f2937,stroke:#2da44e,color:#e6edf3
    style RE fill:#3f1d1d,stroke:#cf3b3b,color:#e6edf3

The loop on the right is the input axis: tool output is billed again on every later request in the session, so shrinking it once pays repeatedly. Read, Edit, and Write are deliberately outside it.

---

Install

One command. Auto-detects your agents (Claude Code, Pi, Cursor, Windsurf, Cline, Kiro, Antigravity, Codex, Gemini, Copilot, Copilot CLI, OpenCode, Hermes) and wires each one. --uninstall puts everything back. npx chisle mcp wrap extends tool-output compression to MCP servers in any agent (details).

npx chisle

# or via curl
curl -fsSL https://raw.githubusercontent.com/JayPokale/Chisle/main/install.sh | bash

# Windows
irm https://raw.githubusercontent.com/JayPokale/Chisle/main/install.ps1 | iex

Preview first with npx chisle --dry-run, scope with --only claude or --only pi, see everything with npx chisle --help. Remove with npx chisle --uninstall.

Upgrading, per-agent setup, --stats and config: docs/usage.md.

---

Benchmarks

Same model, same prompts, one difference: the injected ruleset. Lower = leaner. Chisle is the leanest on every row.

Claude Code (Haiku 4.5, 26 live prompts, tools off, raw cells):

% of a bare model	caveman	ponytail	Chisle
answer length	65%	58%	56%
answer lines	66%	54%	52%

Pi (GPT-5.5, 6 tasks, raw cells):

% of a bare model	caveman	ponytail	Chisle
answer length	56%	49%	37%
answer lines	75%	38%	37%
billed tokens, explanations	43%	76%	39%

Accuracy holds against the bare model: 156/156 vs 155/156 on auto-graded maths and coding, 48/60 vs 49/60 on agentic hidden-test repos, Fisher p = 1.000 on both (writeup). No judge model anywhere.

Tool output is two thirds of a Claude Code session's context and re-billed every turn. The compressor cuts ~46% off every oversized output (171 sessions); on OpenCode a 30 KB log dropped 87% of fresh input tokens.

Method and every raw cell: docs/benchmarks.md.

---

How it works

Before writing code, the agent stops at the first rung that holds:

flowchart TD
    A[Request for code] --> R[Read the problem fully]
    R --> Q1{Does this need<br/>to exist at all?}
    Q1 -->|no| S1[Skip it. Say so in one line]
    Q1 -->|yes| Q2{Already in<br/>this codebase?}
    Q2 -->|yes| S2[Reuse it. Don't rewrite]
    Q2 -->|no| Q3{Stdlib<br/>does it?}
    Q3 -->|yes| S3[Use the stdlib]
    Q3 -->|no| Q4{Native platform<br/>feature covers it?}
    Q4 -->|yes| S4["CSS over JS, DB constraint<br/>over app code"]
    Q4 -->|no| Q5{Already-installed<br/>dependency?}
    Q5 -->|yes| S5[Use it. Never add a new dep<br/>for what a few lines do]
    Q5 -->|no| Q6{Can it be<br/>one line?}
    Q6 -->|yes| S6[One line]
    Q6 -->|no| S7[The minimum code that works]

    S1 & S2 & S3 & S4 & S5 & S6 & S7 --> OUT[Ship it + note what was skipped<br/>and when to add it]

    style Q1 fill:#1f2937,stroke:#d78a3c,color:#e6edf3
    style OUT fill:#14532d,stroke:#2da44e,color:#e6edf3
    style R fill:#1f2937,stroke:#8b949e,color:#e6edf3

The ladder runs after reading, never instead of it. Note the exit: every rung lands on the same obligation: say what you skipped, so "later" doesn't quietly become "never".

The ladder runs after reading the code, lazy about the solution and never about understanding. Lazy is not negligent: trust-boundary validation, data-loss handling, security, and accessibility are never on the chopping block.

Mark deliberate simplifications so "later" doesn't quietly become "never":

// chisle: global lock, per-account locks if throughput matters
// chisle: O(n) scan, index this when table exceeds ~10k rows

Usage

Command	Effect
(nothing)	On automatically every session after install
/chisle	Re-activate if you'd stopped it
/chisle off	Deactivate
stop chisle	Deactivate (ruleset and input-side compression)
normal mode	Deactivate

Natural language works too: "activate chisle", "chisle mode", "chislify this". Code symbols, function/API names, and error strings stay verbatim, so only the noise around them compresses.

---

Contributing

See CONTRIBUTING.md. Built by Jay Pokale with Claude, Antigravity, and Codex as co-engineers.

Star History

License

MIT. The shortest license that works.

Chisle began as a merge of caveman (terse prose) and ponytail (lazy-senior-dev code rules), and its ruleset still carries parts of both. Their MIT notices are in NOTICE.

Contributors

AI co-engineers (pair-work credited in commit trailers and the changelog):

---