ccppccpp home

Profiles

LiteRT-LM Core 0.18.0

Browser integration guardrails for the published @litert-lm/core 0.18.0 package.

Set this profile

ccpp cpo profile set litert-lm-core-018

Inside Claude Code, prefix the command with ! to run it in the session. Downloading a profile needs a Builder or Team plan; listing them does not.

Policies

30 policies, grouped by what they guard.

HTTP & APIs 2

Request and response contracts, schemas, status codes, error bodies and published API surfaces.

  • cli-04-cli-exports

    CLI-04 · LiteRT-LM CLI commands are not JavaScript exports

  • tool-01-preface-tools

    TOOL-01 · Register AutoToolChat tools in its tools option, not config.preface.tools

Dependencies & toolchain 28

Packages change through the package manager, lockfiles stay consistent and the project's tools are used.

  • base-01-version

    BASE-01 · Report a known installed LiteRT-LM version mismatch

  • base-02-python-export

    BASE-02 · Do not import the Python compute_embedding symbol from JavaScript core

  • base-03-public-imports

    BASE-03 · Import LiteRT-LM through its public package entries

  • base-04-commonjs-require

    BASE-04 · Review CommonJS-only loading of the browser ESM package

  • cli-01-dependency-edit

    CLI-01 · CLI-03 · Prefer a package-manager command to hand-editing the LiteRT-LM dependency

  • cli-02-lockfile

    CLI-02 · Let the package manager maintain the LiteRT-LM lockfile resolution

  • emb-01-engine-kind

    EMB-01 · Use the embedding engine for LiteRT-LM embedding calls

  • emb-02-vector-response

    EMB-02 · Do not array-destructure an awaited LiteRT-LM embedding response

  • gen-01-engine-create

    GEN-01 · Await LiteRT-LM Engine.create

  • gen-02-context-budget

    GEN-02 · Review exceptionally large LiteRT-LM context budgets

  • gen-03-conversation-promise

    GEN-03 · Await a LiteRT-LM conversation before invoking its send methods

  • gen-04-system-prompt

    GEN-04 · Use preface messages rather than systemPrompt in LiteRT-LM conversations

  • gen-05-await-message

    GEN-05 · Await LiteRT-LM sendMessage before accessing content

  • gen-06-stream-parts

    GEN-06 · Narrow LiteRT-LM streaming content before indexing text

  • gen-08-overlap

    GEN-08 · Avoid concurrent sends on one LiteRT-LM conversation

  • life-01-await-delete

    LIFE-01 · Await LiteRT-LM conversation deletion

  • life-02-await-delete

    LIFE-02 · Await LiteRT-LM engine deletion

  • life-03-await-delete

    LIFE-03 · Await LiteRT-LM embedding engine deletion

  • mig-01-old-embedding

    MIG-01 · EmbeddingEngine is not a root export of LiteRT-LM 0.15.0

  • mig-03-python-embedding

    MIG-03 · Do not use Python compute_embedding on a JS LiteRT-LM engine

  • sec-01-prompt-secret

    SEC-01 · Keep secret environment values out of LiteRT-LM prompts

  • sec-02-model-html

    SEC-02 · Do not assign unsanitized LiteRT-LM model content to innerHTML

  • tool-02-execute-required

    TOOL-02 · Provide an execute function for each literal AutoToolChat tool

  • tool-07-await-delete

    TOOL-07 · Await AutoToolChat deletion during teardown

  • web-02-localhost-assets

    WEB-02 · Review localhost model URLs before deploying browser inference

  • web-03-model-format

    WEB-03 · Review non-litertlm model files used with LiteRT-LM engines

  • web-04-server-inference

    WEB-04 · Review LiteRT-LM engine creation in explicit server-only modules

  • web-05-insecure-model

    WEB-05 · Review insecure HTTP LiteRT-LM model URLs