mirek

mssqlite system architecture: package dependency graph, request lifecycle from TDS packet to SQLite and back, key design decisions (name flattening, NOCASE collation, parameter passthrough, UDF strategy, session model, metadata inference) and current…

Comprehensive T-SQL language reference for implementing a T-SQL engine. Covers data types, type conversion, operators, expressions, control flow, queries (SELECT, JOIN, CTE, window functions), DML statements (INSERT, UPDATE, DELETE, MERGE), DDL statements…

Comprehensive TDS (Tabular Data Stream) protocol reference for implementing MSSQL clients and servers. Covers packet framing, message types, token streams, data types, login, prelogin, SQL batch, RPC, transactions, error handling, and wire format details. Use…

MSSQL system catalog views technical specification for implementing sys.* emulation in mssqlite. Covers schema definitions, column types, relationships, built-in data, and SQLite backing-table design for all essential catalog views.

Using the tedious MSSQL client against mssqlite — connection options that matter (TLS, port, useColumnNames), Request API and events (row, infoMessage on the connection, done vs doneInProc), how tedious maps operations to the wire (sp_executesql RPC,…

mssqlite repository conventions: pnpm workspace layout, prelude code style (file-per-function, namespace re-exports, immutable readers/results), node-native TypeScript execution, eslint/vitest toolchain, how to add packages and what every package must ship…

Comprehensive SQLite reference for implementing an MSSQL-compatible server on top of SQLite. Covers SQL syntax, type affinity and STRICT tables, DDL (CREATE/ALTER/DROP TABLE/INDEX/VIEW/TRIGGER, FKs, generated columns), DML…

Comprehensive reference for the built-in Node.js node:sqlite module. Covers DatabaseSync, StatementSync, SQLTagStore, Session/changesets, user-defined functions and aggregates, authorization, online backup, serialize/deserialize, type conversion (BigInt and…

CAVE persistence and query specification (§9, §12–§13, §20, §24–§32). Covers append-only belief history, claim identity, SQLite storage, canonicalization, CAVE-Q and temporal reads, EXPECTS shape checks and gates, rules, actions, resolution, aliases, sync,…

CAVE language reference (spec §3–§8, §11, §16, §22) — syntax, lexical rules, verbs and REVERSE inverses, metadata qualifiers, values/units/uncertainty, trajectory values (A -> B) and time contexts, indentation and continuation, tags and topics, normative…

Use CAVE (Compressed Atomic Verb Expressions) as durable, local, append-only agent memory through its cave_* MCP tools or CLI. Use when recording project knowledge, decisions, constraints, evidence, provenance, or confidence; recalling stored facts; querying…

CAVE background and rationale (spec §0–§2, §10, §17–§19) — status conventions, design goals, the claim model and language layers, the probabilistic implementation layer (Bayesian fusion, noisy-AND, competing hypotheses), the Draft unified grammar (variables,…

CAVE extraction guide (spec §14–§15, §21, §23) — rules for converting prose, conversations, code, and structured data into CAVE claims; granularity guide, conversation compaction, boundary cases, operating modes, worked example, deterministic structured…

How to run and observe the cave CLI and MCP server end-to-end when verifying a change in this repo — build-free launch, driving cave commands against a scratch store, and exercising the MCP stdio surface with raw JSON-RPC lines.

Typst 0.15.0 authoring and PDF production guide. Use when creating or editing .typ sources, building books or technical documents, validating generated PDFs, or updating Typst-specific syntax. Version-pinned to the latest stable Typst release verified on…

Develop TypeScript applications with the @prelude/* package family (mirek/prelude). Use when choosing a Prelude package, writing code against Prelude APIs — channels, actors and supervisors, async generators, parser combinators, validation…

How to land a change in mirek/prelude — one PR per fix with a confirming test, green CI, and the automatic chatgpt-codex review awaited and fully resolved before merging. Use whenever creating, updating, or merging a pull request in this repo.

Implement or change mq parser, selector, schema, mutation, library API, or CLI behavior. Use for any code task in the mirek/mq repository that must follow its specification, package boundaries, dependency policy, and lossless-edit invariants.

Design, review, or revise mq syntax and semantics in SPEC.md. Use when deciding Markdown tree behavior, selectors, query expressions, edits, schemas, diagnostics, CLI behavior, compatibility, or implementation sequencing before or alongside code changes.

Implement or change the ast graph model, adapters, query runtime, selectors, change planning, TypeScript API, plugins, or CLI. Use for code tasks in the mirek/ast repository that must follow SPEC.md, the dependency-ordered TODOs, safety invariants, package…

Design, review, or revise ast architecture, public semantics, adapter contracts, selector or DSL behavior, safety rules, and implementation sequencing. Use when changing SPEC.md, decomposing or revising files under todo/, resolving open design questions, or…