Chisle
The coding-agent plugin that bills 52% of a bare model across 20 live tasks. Terse persona + tool-output compressor + context diet.
翻訳待ち
The maker's introduction consists of the line above. Read the repository README for more details. Repository README ↗
Loading the full README.
Read the full README ↗Open measurement SDK for education AI — evidence events, record verification, and dataset clients.
Apache-2.0 · Developer Portal
A typed evidence-event vocabulary, a hash-chained record format, and a verifier that checks bookkeeping integrity without containing any estimation mathematics. The verifier runs locally, makes no network calls, and records never leave your machine.
Not the estimation engine (that runs server-side). Not a learning analytics system. Not a replacement for the measurement service.
Python:
pip install qlm-measure
JavaScript / TypeScript:
npm install qlm-measure
See docs/recorder-quickstart.md — ten lines, no QLM account, no network.
qlm-measure verify record.json
qlm-measure verify records.jsonl --format json --report report.json
qlm-measure explain HASH.ENTRY_RECOMPUTE
qlm-measure version
Exit codes: 0 all valid, 1 at least one invalid, 2 usage or parse error.
Per-event chains sealed by QLM products (TeachProof clinical evidence, Play
clinical events) are defined once, by name and version, in
qlm-measure/schemes — see docs/chain-schemes.md.
qlm-measure verify-chain events.json --family tpc/clinical
qlm-measure verify-chain --list
pip install qlm-measurenpm install qlm-measureBoth packages ship the same 10 checks, the same report schema, and produce byte-identical reports (minus the tool block). The check catalog, canonicalization spec, and golden vectors are shared.
The prepared QLM Transformer and Airflow workflows add assistance definitions, capture declarations, explicit unknowns and native research-export mappings. See the release plan and demo source. These tools are separate from the measurement SDK and hosted estimation engine.
The SDK is Apache-2.0. See LICENSE. Transformer and Airflow subtrees retain their MIT licenses; data and specifications retain separate terms. See license scopes.