# context-vault - Doramagic AI Context Pack

> Positioning: a pre-install experience and judgment asset. It helps the host AI get off to a good start, but it does not mean the project has already been installed, run, or validated.

## Sufficiency Principle

- **Sufficiency over compression**: The AI Context Pack should be sufficient for the host AI to understand the project's value, capability boundaries, entrypoints, risks, and evidence sources before starting work; it may be layered, but it does not aim for the shortest possible summary.
- **Compression policy**: Compress only noise and duplication, never context that affects judgment or the quality of the work.

## How the Host AI Should Use This

You are reading the AI Context Pack that Doramagic compiled for context-vault. Treat it as pre-work context: help the user understand who it fits, what it can do, how to start, what must be verified after install, and where the risks are. Do not claim that you have already installed, run, or executed the target project.

## Claim Consumption Rules

- **Fact source**: Repo Evidence + Claim/Evidence Graph; the Human Wiki only supplies salience, terminology, and narrative structure.
- **Minimum status for a fact**: `supported`
- `supported`: May be used as a project fact, but the answer must cite the claim_id and evidence path.
- `weak`: Usable only as a low-confidence lead; the user must be asked to keep verifying.
- `inferred`: Usable only for risk notes or open questions; must not be packaged as a project fact.
- `unverified`: Must not be used as fact; state clearly that evidence is insufficient.
- `contradicted`: Must show the conflicting sources and must not force a single version on the user's behalf.

## Who It Fits Best

- **Developers already using host AIs such as Claude/Codex/Cursor/Gemini**: The README or plugin config mentions multiple host AIs. Evidence: `README.md` Claim: `clm_0003` supported 0.86

## What It Can Do

- **AI Skill / Agent Instruction Asset Library** (Previewable before install): The project contains Skill or Agent instruction files that a host AI can read, useful for bringing professional workflows into hosts like Claude, Codex, or Cursor. Evidence: `packages/local/assets/skills/compile-context/skill.md`, `packages/local/assets/skills/vault-setup/skill.md` Claim: `clm_0001` supported 0.86
- **Command-Line Startup or Install Flow** (Verify after install): The project documentation contains runnable commands; real use requires running them in a local or host environment. Evidence: `README.md`, `packages/local/README.md`, `packages/sdk/README.md` Claim: `clm_0002` supported 0.86

## How to Start

- `curl -fsSL https://raw.githubusercontent.com/fellanH/context-vault/main/install.sh | sh` Evidence: `README.md` Claim: `clm_0004` supported 0.86
- `git clone https://github.com/fellanH/context-vault.git` Evidence: `README.md` Claim: `clm_0005` supported 0.86
- `npm install -g context-vault` Evidence: `packages/local/README.md` Claim: `clm_0006` supported 0.86
- `npm install @context-vault/sdk` Evidence: `packages/sdk/README.md` Claim: `clm_0007` supported 0.86

## Continue-or-Stop Decision Card

- **Current recommendation**: Sandbox trial only
- **Why**: The project has signals of install commands, host configuration, or local writes; do not go straight into your primary environment—trial it in isolation first.

### 30-Second Read

- **What to do now**: Sandbox trial only
- **Minimum safe next step**: Run Prompt Preview first; if you still want to install, trial only in an isolated environment
- **Do not trust yet**: Real output quality cannot be trusted before install.
- **Continuing will touch**: Command execution, Host AI configuration, Local environment or project files

### What You Can Trust Now

- **Target-audience signal: Developers already using host AIs such as Claude/Codex/Cursor/Gemini** (supported): Backed by a supported claim or project evidence, but that still is not the same as real install results. Evidence: `README.md` Claim: `clm_0003` supported 0.86
- **Capability exists: AI Skill / Agent Instruction Asset Library** (supported): You can trust that the project contains signals of this capability; whether it fits your specific task still needs trial or after-install verification. Evidence: `packages/local/assets/skills/compile-context/skill.md`, `packages/local/assets/skills/vault-setup/skill.md` Claim: `clm_0001` supported 0.86
- **Capability exists: Command-Line Startup or Install Flow** (supported): You can trust that the project contains signals of this capability; whether it fits your specific task still needs trial or after-install verification. Evidence: `README.md`, `packages/local/README.md`, `packages/sdk/README.md` Claim: `clm_0002` supported 0.86
- **There are Quick Start / install-command signals** (supported): You can trust that the docs mention a startup or install entrypoint; do not run it directly in your primary environment because of that. Evidence: `README.md` Claim: `clm_0004` supported 0.86

### What You Cannot Trust Yet

- **Real output quality cannot be trusted before install.** (unverified): Prompt Preview can only show how it guides you; it cannot prove result quality in the real project.
- **Host AI version compatibility cannot be trusted before install.** (unverified): Host loading rules and version differences across Claude, Cursor, Codex, Gemini, and others must be verified in a real environment.
- **That it will not pollute your existing host AI's behavior cannot be trusted directly.** (inferred): Skill, plugin, and AGENTS/CLAUDE/GEMINI instructions may change the host AI's default behavior. Evidence: `packages/local/assets/skills/compile-context/skill.md`, `packages/local/assets/skills/vault-setup/skill.md`
- **Safe rollback cannot be assumed by default.** (unverified): Unless the project clearly provides uninstall and recovery instructions, verify in an isolated environment first.
- **After a real install, is it compatible with the user's current host AI version?** (unverified): Compatibility can only be verified in the actual host environment.
- **Does the project's output quality meet the user's specific task?** (unverified): The pre-install preview can only show flow and boundaries; it cannot replace real evaluation.
- **Do the install commands require network access, permissions, or global writes?** (unverified): This affects install risk in both enterprise and personal environments. Evidence: `README.md`

### What Continuing Will Touch

- **Command execution**: Package managers, network downloads, the local plugin directory, project config, or the user's home directory. Why: Running the very first command can already change your environment; decide whether it is worth running first. Evidence: `README.md`, `packages/local/README.md`, `packages/sdk/README.md`
- **Host AI configuration**: The plugin, Skill, or rule-loading config of hosts like Claude/Codex/Cursor/Gemini/OpenCode. Why: Host configuration changes how the AI works afterward and may conflict with the user's existing rules. Evidence: `packages/local/assets/skills/compile-context/skill.md`, `packages/local/assets/skills/vault-setup/skill.md`
- **Local environment or project files**: Install results, plugin caches, project config, or local dependency directories. Why: The write scope and rollback path cannot be proven before install and need isolated verification. Evidence: `README.md`, `packages/local/README.md`, `packages/sdk/README.md`
- **Host AI context**: The AI Context Pack, Prompt Preview, Skill routing, risk rules, and project facts. Why: Importing context affects the host AI's later judgment, so avoid packaging unverified items as facts.

### Minimum Safe Next Steps

- **Run Prompt Preview first**: Use a pre-install interactive trial to judge whether the way of working fits; it needs no authorization or environment change. (applies when: Applies to any project, especially when output quality is unknown.)
- **Trial-install only in an isolated directory or a test account**: Avoid letting install commands pollute your primary host AI, real projects, or home directory. (applies when: When there are signals of command execution, plugin config, or local writes.)
- **Back up your host AI configuration first**: Skill, plugin, and rule files may change the default behavior of Claude/Cursor/Codex. (applies when: When there is a plugin manifest, a Skill, or a host rule entrypoint.)
- **After install, verify just one minimal task**: Verify loading, compatibility, output quality, and rollback first, then decide whether to use it deeply. (applies when: When moving from a trial into a real workflow.)

### Exit Plan

- **Preserve the pre-install state**: Record the original host config and project state so you can later judge whether it is recoverable.
- **Be ready to remove the host plugin / Skill / rule entrypoint**: If behavior is off after the trial install, you can restore the host AI to its pre-trial state.
- **Record the install commands and written paths**: Without clear uninstall instructions, you at least need to know which directories or configs to clean up manually.
- **If there is no rollback path, do not enter your primary environment**: No rollback is a blocker before continuing; do not proceed on trust or luck.

## What Can Only Be Previewed

- Explain who the project fits and what it can do
- Demonstrate a typical conversation flow based on project docs
- Help the user decide whether it is worth installing or researching further

## What Must Be Verified After Install

- Actually installing the Skill, plugin, or CLI
- Running scripts, modifying local files, or accessing external services
- Verifying real output quality, performance, and compatibility

## Boundary & Risk Decision Card

- **Mistaking the pre-install preview for a real run**: The user may overestimate how much configuration, permission, and compatibility verification the project has already done. Mitigation: Clearly separate prompt_preview_can_do from runtime_required. Claim: `clm_0008` inferred 0.45
- **Command execution will modify the local environment**: Install commands may write to the user's home directory, the host plugin directory, or project configuration. Mitigation: Run in an isolated environment or a test account first. Evidence: `README.md`, `packages/local/README.md`, `packages/sdk/README.md` Claim: `clm_0009` supported 0.86
- **To confirm**: After a real install, is it compatible with the user's current host AI version?. Why: Compatibility can only be verified in the actual host environment.
- **To confirm**: Does the project's output quality meet the user's specific task?. Why: The pre-install preview can only show flow and boundaries; it cannot replace real evaluation.
- **To confirm**: Do the install commands require network access, permissions, or global writes?. Why: This affects install risk in both enterprise and personal environments.

## Pre-Work Working Context

### Loading Order

- First read how_to_use.host_ai_instruction to establish the boundaries of this pre-install judgment asset.
- Read claim_graph_summary to confirm facts come from the Claim/Evidence Graph, not the Human Wiki narrative.
- Then read intended_users, capabilities, and quick_start_candidates to judge whether the user is a match.
- When you need to carry out a concrete task, check role_skill_index first, then evidence_index.
- For real install, file modification, network access, performance, or compatibility questions, turn to risk_card and boundaries.runtime_required.

### Task Routes

- **AI Skill / Agent Instruction Asset Library**: Use role_skill_index / evidence_index to help the user pick a usable role, Skill, or workflow first. Boundary: Can be experienced via a pre-install Prompt. Evidence: `packages/local/assets/skills/compile-context/skill.md`, `packages/local/assets/skills/vault-setup/skill.md` Claim: `clm_0001` supported 0.86
- **Command-Line Startup or Install Flow**: State that this is an after-install capability first, then give a pre-install checklist. Boundary: Must be verified after a real install or run. Evidence: `README.md`, `packages/local/README.md`, `packages/sdk/README.md` Claim: `clm_0002` supported 0.86

### Context Scale

- Total files: 218
- Important-file coverage: 40/218
- Evidence index entries: 80
- Role / Skill entries: 2

### Handling Insufficient Evidence

- **missing_evidence**: State that evidence is insufficient and ask the user for the target file, a README section, or after-install verification records; do not fill in facts.
- **out_of_scope_request**: State that the task is beyond the current AI Context Pack's evidence scope and suggest the user check the Human Manual or verify after a real install.
- **runtime_request**: Provide a pre-install checklist and command sources, but do not run commands for the user or claim they have been run.
- **source_conflict**: Show the conflicting sources side by side, mark them as unverified, and do not force a single version.

## Prompt Recipes

### Fit assessment

- Goal: Judge whether this project fits the user's current task.
- Expected output: A fit conclusion, key reasons, evidence citations, what can be previewed before install, what must be verified after install, and a next-step recommendation.

```text
Based on the AI Context Pack for context-vault, ask me 3 necessary questions first, then judge whether it fits my task. The answer must cover: who it fits, what it can do, what it cannot do, whether it is worth installing, and where the evidence comes from. Every project fact must cite evidence_refs, source_paths, or a claim_id.
```

### Pre-install experience

- Goal: Let the user feel the core workflow before installing, while avoiding packaging the preview as real capability or a marketing promise.
- Expected output: An experience script with boundary labels, an after-install verification checklist, and a cautious recommendation; with no real-run promises or strong marketing language.

```text
Treat context-vault as a pre-install experience asset, not an already-installed tool or a real runtime environment.

Output exactly four parts:
1. Ask me 3 necessary questions first.
2. Give an "experience script": use the three labels [Previewable before install], [Must verify after install], and [Insufficient evidence] to show how it might guide the workflow.
3. Give an after-install verification checklist: list which capabilities can only be confirmed after a real install, real host loading, and a real project run.
4. Give a cautious recommendation: only "worth researching/trialing further", "add information before deciding", or "not recommended to continue"; do not endorse the project.

Hard boundaries:
- Do not claim you have installed, run, executed tests, modified files, or produced real results.
- Do not write promise-like phrasing such as "auto-adapts", "guarantees passing", "perfect fit", or "strongly recommend installing".
- If you describe how it works after install, you must use a conditional such as "if installed successfully and the host loads the Skill correctly, it might...".
- The experience script may only be written as "example lines / hypothetical flow": use "might ask / might suggest / might show", not "has written, has generated, has passed, is running, is generating".
- Prompt Preview does not hand out install commands; if the user is ready to trial, only prompt them to read Quick Start and the Risk Card first and to verify in an isolated environment.
- Every project fact must come from a supported claim, evidence_refs, or source_paths; inferred/unverified items can only be risks or open questions.

```

### Role / Skill selection

- Goal: Pick the best-matching asset from the project's roles or Skills.
- Expected output: A list of candidate roles or Skills, each with an applicable scenario, evidence paths, risk boundary, and whether after-install verification is needed.

```text
Read role_skill_index and recommend 3-5 of the most relevant roles or Skills for my target task. For each recommendation, state the applicable scenario, likely output, risk boundary, and evidence_refs.
```

### Risk pre-check

- Goal: Identify environment, permission, rule-conflict, and quality risks before installing or adopting.
- Expected output: A checklist of environment, permission, dependency, license, host-conflict, quality risk, and unknown items.

```text
Based on risk_card, boundaries, and quick_start_candidates, give me a pre-install risk pre-check list. Do not run commands for me; only explain what I should check, why, and what impact a failure would have.
```

### Host AI kickoff instruction

- Goal: Turn the project context into a host AI instruction for the start of a conversation.
- Expected output: A pre-work instruction with clear boundaries and clear evidence citations, suitable to copy to a host AI.

```text
Based on the AI Context Pack for context-vault, generate a pre-work instruction I can paste to my host AI. This instruction must obey not_runtime=true and must not claim the project has been installed, run, or produced real results.
```

## Role / Skill Index

- Indexed 2 role / Skill / project-doc entries.

- **compile-context** (skill):  Activation hint: When the user's task is highly relevant to the workflow described by “compile-context”, use it for a pre-install experience first, then decide whether to install. Evidence: `packages/local/assets/skills/compile-context/skill.md`
- **vault-setup** (skill):  Activation hint: When the user's task is highly relevant to the workflow described by “vault-setup”, use it for a pre-install experience first, then decide whether to install. Evidence: `packages/local/assets/skills/vault-setup/skill.md`

## Evidence Index

- Indexed 80 evidence entries.

- **context-vault** (documentation): ! npm version https://img.shields.io/npm/v/context-vault https://www.npmjs.com/package/context-vault ! license https://img.shields.io/npm/l/context-vault ./LICENSE ! node https://img.shields.io/node/v/context-vault https://nodejs.org Evidence: `README.md`
- **context-vault** (documentation): ! npm version https://img.shields.io/npm/v/context-vault https://www.npmjs.com/package/context-vault ! npm downloads https://img.shields.io/npm/dm/context-vault https://www.npmjs.com/package/context-vault ! license https://img.shields.io/npm/l/context-vault ./LICENSE ! node https://img.shields.io/node/v/context-vault https://nodejs.org Evidence: `packages/local/README.md`
- **@context-vault/sdk** (documentation): Context lifecycle SDK for AI agents. Save, search, and manage persistent context with freshness scoring. Evidence: `packages/sdk/README.md`
- **Package** (package_manifest): { "name": "context-vault-monorepo", "private": true, "version": "4.0.0", "type": "module", "workspaces": "packages/local", "packages/sdk" , "scripts": { "build": "cargo build --release --workspace", "cli": "node packages/local/bin/cli.js", "test": "cargo test --workspace", "release": "node scripts/release.mjs" }, "engines": { "node": " =18" }, "devDependencies": { "typescript": "^5.7.0" } } Evidence: `package.json`
- **Package** (package_manifest): { "name": "context-vault", "version": "4.0.0", "type": "module", "description": "Persistent memory for AI agents — saves and searches knowledge across sessions", "bin": { "context-vault": "bin/cli.js", "context-mcp": "bin/cli.js", "memory": "bin/cli.js" }, "exports": { "./cli": "./bin/cli.js" }, "scripts": {}, "files": "assets/", "bin/", ".mcp.json", "README.md", "LICENSE" , "license": "MIT", "engines": { "node": " =18" }, "author": "Felix Hellstrom", "repository": { "type": "git", "url": "git+https://github.com/fellanH/context-vault.git" }, "homepage": "https://github.com/fellanH/context-vault", "bugs": { "url": "https://github.com/fellanH/context-vault/issues" }, "keywords": "mcp", "model… Evidence: `packages/local/package.json`
- **Package** (package_manifest): { "name": "@context-vault/sdk", "version": "0.1.0", "description": "Context lifecycle SDK for AI agents. Save, search, and manage persistent context with freshness scoring.", "type": "module", "main": "dist/index.js", "types": "dist/index.d.ts", "exports": { ".": { "types": "./dist/index.d.ts", "import": "./dist/index.js" }, "./hosted": { "types": "./dist/hosted.d.ts", "import": "./dist/hosted.js" } }, "files": "dist" , "license": "Apache-2.0", "keywords": "context", "ai", "agents", "memory", "mcp", "lifecycle" , "devDependencies": { "@types/node": "^22.0.0", "typescript": "^5.5.0" } } Evidence: `packages/sdk/package.json`
- **compile-context skill** (skill_instruction): When the user asks to compile context or create a brief for a topic, call create snapshot to synthesize a context brief from the vault. Evidence: `packages/local/assets/skills/compile-context/skill.md`
- **vault-setup skill** (skill_instruction): Agent-assisted setup that customizes the context-vault installation for the user's workflow. Always show what you plan to do before doing it. Never modify existing rules without permission. Evidence: `packages/local/assets/skills/vault-setup/skill.md`
- **License** (source_file): This repository uses a dual-license model: Evidence: `LICENSE`
- **License** (source_file): Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files the "Software" , to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: Evidence: `packages/local/LICENSE`
- **--vault-db FTS probe** (documentation): Rust context-vault bench --vault-db runs read-only probes. FTS token selection is not the fixture hybrid path; see timings.sqlite.fts user probe in JSON output. Evidence: `docs/bench-vault-fts.md`
- **Claude Code Plugin** (documentation): Context Vault integrates with Claude Code via a UserPromptSubmit hook. On every prompt, it searches your vault for relevant entries and injects them as a XML block that Claude receives alongside your message. Evidence: `docs/claude-code-plugin.md`
- **Development Workflow** (documentation): How to make changes to context-vault, release them, and verify they work as a real user would experience. Evidence: `docs/development-workflow.md`
- **Rust binary distribution** (documentation): See specs/rust-core.md for engine design and docs/npm-rust-migration.md for the migration stages. This file covers the binary packaging and release mechanics only. Evidence: `docs/rust-distribution.md`
- **npm → Rust CLI migration** (documentation): Archived 2026-07-03. The actual cutover skipped the staged Stage-1-IPC approach below and went straight to deleting the TS engine 2026-07-02 . Current status lives in project-management/docs/rust-cutover-runbook.md . Kept for historical context. Evidence: `docs/archive/npm-rust-migration.md`
- **sqlite-vec hooks** (documentation): Archived 2026-07-03. Resolved — sqlite-vec does load via rusqlite 's bundled feature; it's wired into crates/cv-backends/src/sqlite.rs and sqlite ops.rs . Kept for historical context on the original open question. Evidence: `docs/archive/rust-sqlite-vec-hooks.md`
- **Connect in 2 Minutes Hosted** (documentation): Use Context Vault in the cloud — no local install, no Node.js. Works across devices. Sign up at app.context-vault.com https://app.context-vault.com , get an API key, then add the MCP endpoint to your AI tool. Evidence: `docs/distribution/connect-in-2-minutes.md`
- **Hosted-Omni Migration Prototype Campaign 2026-06-17** (documentation): Hosted-Omni Migration Prototype Campaign 2026-06-17 Evidence: `project-management/docs/2026-06-17-hosted-omni-prototype-campaign.md`
- **Rust cutover runbook — Phase 7** (documentation): Status updated 2026-07-03 : Engine migration Phases 0–6 and Phase 7's TS deletion are COMPLETE, committed, and pushed to origin/main . The session-end parity gap below is closed. This dev machine's global context-vault is now the native Rust binary , installed directly via cargo install --path crates/context-vault-cli no Node in the loop at all — the old npm package was uninstalled . The outward-facing public npm publish is a separate, NOT-yet-executed step — it ships to a live npm product via a cross-compile path that has never actually run, and needs human sign-off see "Remaining: the actual release" below . Evidence: `project-management/docs/rust-cutover-runbook.md`
- **Shared dependency versions. Members opt in via .workspace = true .** (source_file): workspace resolver = "2" members = "crates/cv-engine", "crates/cv-vault", "crates/cv-schema", "crates/cv-lifecycle", "crates/cv-ingest", "crates/cv-backends", "crates/context-vault-core", "crates/context-vault-cli", "crates/context-vault-mcp", "crates/context-vault-gateway", Evidence: `Cargo.toml`
- **bench + invoke are maintainer/debug tooling, not part of the shipped command surface —** (source_file): package name = "context-vault-cli" version = "4.0.0" edition.workspace = true license.workspace = true description = "Context Vault CLI — engine subcommands mcp, save/get/search/list/delete, lifecycle . Ships the context-vault binary." Evidence: `crates/context-vault-cli/Cargo.toml`
- **Client Paths** (source_file): use std::path::PathBuf; ⋮---- pub fn get app data dir - PathBuf { ⋮---- .expect "home dir" .join "Library/Application Support" ⋮---- .map PathBuf::from .unwrap or else { ⋮---- .join "AppData/Roaming" ⋮---- .join ".local/share" ⋮---- pub fn get vscode data dir - PathBuf { ⋮---- .join "Library/Application Support/Code/User/globalStorage" ⋮---- .map p PathBuf::from p .join "Code/User/globalStorage" ⋮---- .join "AppData/Roaming/Code/User/globalStorage" ⋮---- .join ".config/Code/User/globalStorage" Evidence: `crates/context-vault-cli/src/client_paths.rs`
- **Main** (source_file): mod engine cmd; mod cmd read; mod cmd write; mod cmd lifecycle; mod cmd ingest; mod cmd setup; mod cmd schema; mod cmd move; mod cmd session end; mod client paths; mod setup register; mod cli; ⋮---- mod ipc invoke; ⋮---- use clap::Parser; ⋮---- use std::path::PathBuf; ⋮---- use anyhow::anyhow; ⋮---- use context vault core::bench harness; ⋮---- use anyhow::Result; ⋮---- async fn main - Result { ⋮---- println! ⋮---- Ok ⋮---- Some cmd = dispatch cmd .await, ⋮---- async fn dispatch cmd: Commands - Result { ⋮---- } = run bench ⋮---- Commands::PostToolCall = Ok , ⋮---- fn run bench ⋮---- return Err anyhow! "only --format json is supported in this slice" ; ⋮---- let mut stdout = stdout.lock ; ⋮---… Evidence: `crates/context-vault-cli/src/main.rs`
- **async-trait, tempfile, and rusqlite bundled are normal deps above and cover the test targets too.** (source_file): package name = "context-vault-core" version = "0.2.2" edition.workspace = true license.workspace = true description = "Context Vault engine: SQLite storage, hybrid FTS+vector search, embeddings, and the markdown vault Rust core library ." Evidence: `crates/context-vault-core/Cargo.toml`
- **Mod** (source_file): mod corpus; pub mod fts token; pub crate mod sqlite lane; ⋮---- use std::fs; use std::path::PathBuf; use std::time::Instant; ⋮---- use serde::Serialize; ⋮---- pub use fts token::FtsPickResult; pub use sqlite lane::SqliteLane; ⋮---- pub enum DiskMode { ⋮---- pub enum WorkloadMix { ⋮---- pub struct BenchArgs { ⋮---- pub struct BenchOutput { ⋮---- pub struct BenchMeta { ⋮---- pub struct Timings { ⋮---- pub struct DiskModesDoc { ⋮---- pub struct WorkloadMixesDoc { ⋮---- pub fn run bench args: BenchArgs - anyhow::Result { ⋮---- let corpus = generate corpus args.corpus root.clone , ⋮---- let corpus generation ms = t0.elapsed .as secs f64 1000.0; ⋮---- sqlite lane::run fixture probes &conn, args.w… Evidence: `crates/context-vault-core/src/bench_harness/mod.rs`
- **Lib** (source_file): pub mod bench harness; ⋮---- pub use cv schema::indexing; ⋮---- pub mod mcp; pub mod paths; Evidence: `crates/context-vault-core/src/lib.rs`
- **Handlers Ingest** (source_file): use std::sync::Arc; use serde json::Value; ⋮---- pub async fn handle ingest url ⋮---- .get "url" .and then v v.as str .ok or "Required: url" ?; ⋮---- if url.len 2048 { return Err "url must be under 2048 chars".to string ; ⋮---- .get "kind" ⋮---- .unwrap or "reference" .to string ; ⋮---- let tags = args.get "tags" .and then v v.as array .map arr { arr.iter .filter map v v.as str .map s s.to string ⋮---- kind: Some kind , ⋮---- let result = ingest url url, opts .await.map err e e.to string ?; ⋮---- title: Some result.title.clone , ⋮---- tags: Some result.tags , source: Some result.source , meta: Some serde json::json! result.meta , ⋮---- let output = engine.save input .await.map err e e.to st… Evidence: `crates/context-vault-core/src/mcp/handlers_ingest.rs`
- **Handlers Read** (source_file): use std::sync::Arc; use serde json::Value; use crate::engine::VaultEngine; ⋮---- pub async fn handle get context ⋮---- let query = args.get "query" .and then v v.as str ; ⋮---- .get "limit" .and then v v.as u64 .unwrap or 10 as usize; ⋮---- .hybrid search query.unwrap or "" , opts ⋮---- .map err e e.to string ?; ⋮---- if results.is empty { return Ok format! "No results found{}", ⋮---- let mut lines = vec! format! " Results {} matches \n", results.len ; ⋮---- let title = e.title.as deref .unwrap or " untitled " ; lines.push format! ⋮---- if let Some body = &e.body.lines .next { lines.push format! " {}", body ; ⋮---- Ok lines.join "\n" ⋮---- pub async fn handle list context ⋮---- let kind = a… Evidence: `crates/context-vault-core/src/mcp/handlers_read.rs`
- **Handlers Write** (source_file): use std::sync::Arc; use serde json::Value; ⋮---- pub async fn handle save context ⋮---- .get "kind" .and then v v.as str .ok or "Required: kind" ? .to string ; ⋮---- .get "body" ⋮---- .ok or "Required: body" ? ⋮---- let title = args.get "title" .and then v v.as str .map s s.to string ; let id = args.get "id" .and then v v.as str .map s s.to string ; ⋮---- let tags = args.get "tags" .and then v v.as array .map arr { arr.iter .filter map v v.as str .map s s.to string .collect ⋮---- meta: args.get "meta" .cloned , source: args.get "source" .and then v v.as str .map s s.to string , identity key: args.get "identity key" .and then v v.as str .map s s.to string , tier: args.get "tier" .and then v… Evidence: `crates/context-vault-core/src/mcp/handlers_write.rs`
- **Mod** (source_file): pub mod stdio; pub mod tools; ⋮---- pub super mod tool schemas; pub super mod handlers read; pub super mod handlers write; pub super mod handlers ingest; ⋮---- pub use stdio::run; ⋮---- use std::sync::Arc; ⋮---- use crate::engine::VaultEngine; ⋮---- pub async fn recall Evidence: `crates/context-vault-core/src/mcp/mod.rs`
- **Stdio** (source_file): use std::sync::Arc; ⋮---- use super::tools; ⋮---- pub async fn run - Result { let db path = resolve db path ?; let vault dir = resolve vault dir ?; let embedder = init embedder .await?; ⋮---- let engine: Arc = backend.clone ; ⋮---- engine.init .await?; ⋮---- let reindex engine = engine.clone ; ⋮---- if let Err e = reindex engine.reindex opts .await { eprintln! " context-vault Background reindex failed: {}", e ; ⋮---- let mut reader = BufReader::new stdin.lock ; let mut stdout = io::stdout .lock ; ⋮---- while process one message &mut reader, &mut stdout, &engine .await? {} ⋮---- Ok ⋮---- async fn process one message ⋮---- let body = match read mcp message reader ? { ⋮---- None = return Ok fa… Evidence: `crates/context-vault-core/src/mcp/stdio.rs`
- **Mod** (source_file): mod core; mod extra; Evidence: `crates/context-vault-core/src/mcp/tool_schemas/mod.rs`
- **Tools** (source_file): use std::sync::Arc; ⋮---- use crate::engine::VaultEngine; ⋮---- pub crate async fn dispatch tool call ⋮---- let params = msg.get "params" .cloned .unwrap or json! {} ; ⋮---- .get "name" .and then n n.as str .unwrap or "unknown" ; let arguments = params.get "arguments" .cloned .unwrap or json! {} ; ⋮---- handle tool engine.clone , name, arguments , ⋮---- vec! json! { "type": "text", "text": text } , false ⋮---- vec! json! { "type": "text", "text": format! "Tool error: {e}" } , true ⋮---- vec! json! { "type": "text", "text": format! "Tool timeout after {}ms", TOOL TIMEOUT MS } , true ⋮---- json! { ⋮---- pub crate fn reply initialize id: Option - Value { ⋮---- pub crate fn reply tools list id:… Evidence: `crates/context-vault-core/src/mcp/tools.rs`
- **Paths** (source_file): use std::path::PathBuf; ⋮---- use serde json::Value; ⋮---- fn data dir path home: &str - String { ⋮---- .or else std::env::var "CONTEXT MCP DATA DIR" .unwrap or else PathBuf::from home .join ".context-mcp" .to string lossy .to string ⋮---- fn read config json home: &str - Option { let path = PathBuf::from data dir path home .join "config.json" ; let contents = std::fs::read to string path .ok ?; serde json::from str &contents .ok ⋮---- pub fn resolve db path - Result { let home = std::env::var "HOME" .context "HOME not set" ?; ⋮---- std::env::var "CONTEXT VAULT DB PATH" .or else std::env::var "CONTEXT MCP DB PATH" ⋮---- read config json &home .and then c c.get "dbPath" ?.as str .map String:… Evidence: `crates/context-vault-core/src/paths.rs`
- **Cargo** (source_file): package name = "context-vault-gateway" version = "0.2.2" edition.workspace = true license.workspace = true description = "Context Vault self-hosted HTTP gateway: a thin axum server exposing the Rust engine over a bearer-authenticated REST API." Evidence: `crates/context-vault-gateway/Cargo.toml`
- **Main** (source_file): mod handlers; ⋮---- use std::sync::Arc; ⋮---- pub crate struct AppState { ⋮---- async fn main - Result { ⋮---- .context "GATEWAY API KEY is required bearer token for the gateway " ?; let host = std::env::var "GATEWAY HOST" .unwrap or else "127.0.0.1".to string ; let port = std::env::var "GATEWAY PORT" .unwrap or else "8787".to string ; ⋮---- let engine = build engine .await?; ⋮---- .route "/api/vault/status", get handlers::status handler .route "/api/vault/entries", post handlers::save handler .route "/api/vault/entries/{id}", get handlers::get handler .delete handlers::delete handler .route "/api/vault/search", get handlers::search handler .route layer from fn with state state.clone , hand… Evidence: `crates/context-vault-gateway/src/main.rs`
- **Cargo** (source_file): package name = "context-vault-mcp" version = "0.2.2" edition.workspace = true license.workspace = true description = "Context Vault MCP stdio server — thin binary over context-vault-core." Evidence: `crates/context-vault-mcp/Cargo.toml`
- **Main** (source_file): use anyhow::Result; ⋮---- async fn main - Result { Evidence: `crates/context-vault-mcp/src/main.rs`
- **Cargo** (source_file): package name = "cv-backends" version = "0.2.2" edition.workspace = true license.workspace = true description = "Context Vault SQLite backend: storage, FTS+vec search, embeddings." Evidence: `crates/cv-backends/Cargo.toml`
- **Embedder** (source_file): use cv engine::Embedder; use async trait::async trait; use std::sync::Arc; ⋮---- pub struct NoOpEmbedder; ⋮---- impl Embedder for NoOpEmbedder { async fn embed &self, text: &str - anyhow::Result { Ok vec! 0.0; 384 ⋮---- fn dim &self - usize { ⋮---- pub struct FastembedEmbedder { ⋮---- impl FastembedEmbedder { pub fn new - anyhow::Result { ⋮---- Ok Self { model: Arc::new model } ⋮---- impl Embedder for FastembedEmbedder { async fn embed &self, text: &str - anyhow::Result { let text = text.to string ; let model = self.model.clone ; ⋮---- let mut results = model.embed vec! text.as str , None ?; results.pop .ok or else anyhow::anyhow! "empty embedding result" Evidence: `crates/cv-backends/src/embedder.rs`
- **Lib** (source_file): pub mod embedder; pub mod queries; pub mod row map; pub mod sqlite; pub mod sqlite ops; ⋮---- pub use sqlite::SqliteBackend; Evidence: `crates/cv-backends/src/lib.rs`
- **Sqlite** (source_file): use async trait::async trait; use anyhow::Result; use rusqlite::Connection; use std::path::PathBuf; ⋮---- use super::queries::entry from row; use super::sqlite ops; ⋮---- fn build list query opts: &ListOptions - String { let mut where clauses = vec! ; ⋮---- where clauses.push format! "kind = '{}'", kind.replace '\'', "''" ; ⋮---- where clauses.push format! "category = '{}'", category.replace '\'', "''" ; ⋮---- where clauses.push format! "tier = '{}'", tier.replace '\'', "''" ; ⋮---- let mut query = LIST COLUMNS.to string ; if !where clauses.is empty { query.push str &format! " WHERE {}", where clauses.join " AND " ; ⋮---- query.push str &format! ⋮---- pub struct SqliteBackend { ⋮---- impl S… Evidence: `crates/cv-backends/src/sqlite.rs`
- **Sqlite Ops** (source_file): use cv vault::capture::WriteResult; ⋮---- use anyhow::Result; use chrono::Utc; use rusqlite::ffi::sqlite3 auto extension; ⋮---- use sqlite vec::sqlite3 vec init; use std::path::Path; ⋮---- use std::collections::HashMap; use ulid::Ulid; use zerocopy::AsBytes; ⋮---- pub super fn register sqlite vec { VEC INIT.call once unsafe { ⋮---- sqlite3 auto extension Some std::mem::transmute ⋮---- pub super fn is zero embedding embedding: & f32 - bool { embedding.is empty embedding.iter .all &x x == 0.0 ⋮---- pub async fn save entry ⋮---- let tags = input.tags.as ref .map t serde json::to string t .unwrap or default ; let meta = input.meta.as ref .map m serde json::to string m .unwrap or default ; ⋮----… Evidence: `crates/cv-backends/src/sqlite_ops.rs`
- **Cargo** (source_file): package name = "cv-engine" version = "0.2.2" edition.workspace = true license.workspace = true description = "Context Vault engine contract: VaultEngine/Embedder traits, data types, and pure RRF search scoring." Evidence: `crates/cv-engine/Cargo.toml`
- **Lib** (source_file): use async trait::async trait; ⋮---- use serde json::Value; ⋮---- pub mod search; ⋮---- pub struct Entry { ⋮---- pub struct SaveInput { ⋮---- pub struct SaveOutput { ⋮---- pub struct SearchOptions { ⋮---- pub struct ScoredEntry { ⋮---- pub struct ListOptions { ⋮---- pub struct ReindexOptions { ⋮---- pub struct ReindexSummary { ⋮---- pub struct VaultStats { ⋮---- pub struct HealthStatus { ⋮---- pub trait Embedder: Send + Sync { ⋮---- pub trait VaultEngine: Send + Sync { Evidence: `crates/cv-engine/src/lib.rs`
- **Cargo** (source_file): package name = "cv-ingest" version = "0.2.2" edition.workspace = true license.workspace = true description = "Context Vault ingest URL/project and file watcher." Evidence: `crates/cv-ingest/Cargo.toml`
- **Mod** (source_file): pub mod url convert; pub mod url; pub mod project; mod session signal extract; pub mod session transcript; pub mod session insights; Evidence: `crates/cv-ingest/src/ingest/mod.rs`
- **Project** (source_file): use std::fs; use std::path::Path; use std::process::Command; ⋮---- mod tech stack; use tech stack::detect tech stack; ⋮---- fn safe read path: &str - Option { fs::read to string path .ok ⋮---- fn safe exec cmd: &str, cwd: &str - Option { let parts: Vec = cmd.split whitespace .collect ; let cmd name = parts.first ?; ⋮---- .args &parts 1.. .current dir cwd .output .ok .and then output { output.status.success .then { ⋮---- .map s s.trim .to string ⋮---- fn extract readme description project path: &str - Option { let readme path = format! "{}/README.md", project path ; let readme lower = format! "{}/readme.md", project path ; let content = safe read &readme path .or else safe read &readme lower… Evidence: `crates/cv-ingest/src/ingest/project.rs`
- **Url** (source_file): use std::time::Duration; use crate::ingest::url convert::extract html content; ⋮---- pub struct IngestResult { ⋮---- pub async fn ingest url ⋮---- let kind = opts.kind.unwrap or else "reference".to string ; let tags = opts.tags.unwrap or default ; ⋮---- let max body length = opts.max body length.unwrap or 50000 ; let timeout ms = opts.timeout ms.unwrap or 15000 ; ⋮---- .map err e anyhow! "Invalid URL: {}", e ?; ⋮---- .host str .ok or else anyhow! "Invalid URL: no host" ? .to string ; ⋮---- .timeout Duration::from millis timeout ms .user agent "ContextVault/1.0 +https://github.com/fellanH/context-vault " .build ?; ⋮---- let response = client.get url .send .await .map err e anyhow! "Fetch fai… Evidence: `crates/cv-ingest/src/ingest/url.rs`
- **Url Convert** (source_file): use regex::Regex; ⋮---- mod html format; ⋮---- pub fn strip tags html: &str - String { ⋮---- .unwrap .replace all html, "" .to string ⋮---- pub fn decode entities text: &str - String { ⋮---- .replace "&", "&" .replace "<", " " .replace """, "\"" .replace "& 39;", "'" .replace " ", " " ; ⋮---- .replace all &result, caps: &regex::Captures { ⋮---- .ok .and then char::from u32 .map ch ch.to string .unwrap or else caps 0 .to string ⋮---- .to string ; ⋮---- pub fn html to markdown html: &str - String { let mut md = html.to string ; ⋮---- // Remove unwanted containers md = remove container &md, "script" ; md = remove container &md, "style" ; ⋮---- md = remove container &md, tag ;… Evidence: `crates/cv-ingest/src/ingest/url_convert.rs`
- **Url Convert Html** (source_file): use regex::Regex; ⋮---- pub super fn remove container md: &str, tag: &str - String { let pattern = format! r" ?i ", tag ; ⋮---- .unwrap .replace all md, "" .to string ⋮---- pub super fn apply heading levels mut md: String - String { ⋮---- let pattern = format! r" ?i \s\S ? ", level ; let hashes = " ".repeat level ; ⋮---- .replace all &md, caps: &regex::Captures { format! "\n{} {}\n", hashes, strip tags &caps 1 .trim ⋮---- .to string ; ⋮---- pub super fn apply link formatting md: &str - String { ⋮---- .replace all md, caps: &regex::Captures { let text str = strip tags &caps 2 ; let text = text str.trim ; if text.is empty { ⋮---- format! " {} {} ", text, &caps 1 ⋮---- pub super fn apply code… Evidence: `crates/cv-ingest/src/ingest/url_convert_html.rs`
- **Lib** (source_file): pub mod ingest; pub mod watch; Evidence: `crates/cv-ingest/src/lib.rs`
- **Mod** (source_file): pub mod watcher; Evidence: `crates/cv-ingest/src/watch/mod.rs`
- **Mod** (source_file): use anyhow::Result; ⋮---- use std::time::SystemTime; ⋮---- mod lock; mod self write; ⋮---- use lock::WatcherLock; use self write::SelfWriteTracker; ⋮---- fn current time ms - u64 { ⋮---- .duration since SystemTime::UNIX EPOCH .map d d.as millis as u64 .unwrap or 0 ⋮---- fn is excluded rel path: &str - bool { if rel path.starts with '.' { ⋮---- .split '/' .any part part.starts with '.' excluded dirs.contains &part ⋮---- pub struct VaultWatcher { ⋮---- impl VaultWatcher { pub fn mark self write &self, file path: String { if let Ok mut tracker = self.self writes.lock { tracker.mark file path ; ⋮---- pub fn should process event &self, rel path: &str - bool { if is excluded rel path { ⋮---- if t… Evidence: `crates/cv-ingest/src/watch/watcher/mod.rs`
- **Cargo** (source_file): package name = "cv-lifecycle" version = "0.2.2" edition.workspace = true license.workspace = true description = "Context Vault lifecycle: summarize, compact, archive, assemble, consolidation, tier analysis." Evidence: `crates/cv-lifecycle/Cargo.toml`
- **Lib** (source_file): pub mod summarize; pub mod compact; pub mod archive; pub mod assemble; pub mod consolidation; pub mod tier analysis; Evidence: `crates/cv-lifecycle/src/lib.rs`
- **Cargo** (source_file): package name = "cv-schema" version = "0.2.2" edition.workspace = true license.workspace = true description = "Context Vault schema + multi-vault registry + indexing policy." Evidence: `crates/cv-schema/Cargo.toml`
- **Indexing** (source_file): pub struct IndexingConfig { ⋮---- impl Default for IndexingConfig { fn default - Self { ⋮---- exclude kinds: vec! , exclude categories: vec! , ⋮---- pub fn should index ⋮---- let c = config.unwrap or &default ; ⋮---- if !c.exclude kinds.is empty && c.exclude kinds.contains &kind.to string { ⋮---- if !c.exclude categories.is empty && c.exclude categories.contains &category.to string { ⋮---- mod tests; ⋮---- mod more tests; Evidence: `crates/cv-schema/src/indexing.rs`
- **Lib** (source_file): pub mod vaults; ⋮---- pub mod indexing; Evidence: `crates/cv-schema/src/lib.rs`
- **Vaults** (source_file): use std::collections::BTreeMap; ⋮---- use anyhow::Result; ⋮---- pub struct VaultEntry { ⋮---- pub struct VaultRegistry { ⋮---- pub fn registry path - PathBuf { ⋮---- .or else std::env::var "CONTEXT VAULT DATA DIR" .unwrap or else { ⋮---- .unwrap or else PathBuf::from "." .join ".context-mcp" .to string lossy .to string ⋮---- Path::new &data dir .join "vaults.json" ⋮---- pub fn load registry - Result { let path = registry path ; ⋮---- if !path.exists { return Ok VaultRegistry { active vault: "default".to string , ⋮---- Ok registry ⋮---- pub fn save registry registry: &VaultRegistry - Result { ⋮---- std::fs::create dir all path.parent .unwrap ?; ⋮---- Ok ⋮---- pub fn resolve active vault regi… Evidence: `crates/cv-schema/src/vaults.rs`
- The remaining 20 evidence entries are in `AI_CONTEXT_PACK.json` or `EVIDENCE_INDEX.json`.

## Rules the Host AI Must Follow

- **Treat this asset as pre-work context, not a runtime environment.**: The AI Context Pack contains only an evidence-backed understanding of the project, not the project's executable state. Evidence: `README.md`, `packages/local/README.md`, `packages/sdk/README.md`
- **When answering the user, distinguish what can be previewed from what can only be verified after install.**: The consumer value of the pre-install experience comes from reducing bad installs and misjudgments, not from pretending to be a real run. Evidence: `README.md`, `packages/local/README.md`, `packages/sdk/README.md`

## Questions the User Should Answer First

- Which host AI or local environment do you plan to use it in?
- Do you just want to experience the workflow first, or are you ready to actually install?
- What matters most to you: install cost, output quality, or conflicts with your existing rules?

## Acceptance Checks

- Every capability claim can be traced back to a file path in evidence_refs.
- AI_CONTEXT_PACK.md does not package previews as a real run.
- The user can understand who it fits, what it can do, how to start, and the risk boundaries within 3 minutes.

---

## Doramagic Context Augmentation

The following sections strengthen the repository context for a host AI. Human Manual data is a reading route, and pitfall notes become operating constraints.

## Human Manual Outline

Usage rule: this is only a reading route and salience signal, not factual authority. Concrete claims must still return to repo evidence or Claim Graph.

Host AI hard rules:
- Do not treat page titles, section order, summaries, or importance values as factual project evidence.
- When explaining the Human Manual outline, state that it is only a reading route or salience signal.
- Capability, installation, compatibility, runtime state, and risk claims must cite repo evidence, source paths, or Claim Graph.

- **System Overview & Workspace Architecture**: importance `high`
  - source_paths: README.md, Cargo.toml, crates/context-vault-core/src/lib.rs, crates/context-vault-cli/src/main.rs, crates/context-vault-core/src/paths.rs
- **MCP Tools, CLI & SDK Surface**: importance `high`
  - source_paths: crates/context-vault-core/src/mcp/mod.rs, crates/context-vault-core/src/mcp/stdio.rs, crates/context-vault-core/src/mcp/tools.rs, crates/context-vault-core/src/mcp/handlers_read.rs, crates/context-vault-core/src/mcp/handlers_write.rs
- **Storage, Indexing & Hybrid Search Engine**: importance `high`
  - source_paths: crates/cv-schema/src/lib.rs, crates/cv-schema/src/indexing.rs, crates/cv-schema/src/vaults.rs, crates/cv-backends/src/sqlite.rs, crates/cv-backends/src/sqlite_ops.rs
- **Ingestion, Lifecycle, Operations & Known Failure Modes**: importance `high`
  - source_paths: crates/cv-ingest/src/lib.rs, crates/cv-ingest/src/ingest/mod.rs, crates/cv-ingest/src/ingest/url.rs, crates/cv-ingest/src/ingest/url_convert.rs, crates/cv-ingest/src/ingest/url_convert_html.rs

## Repo Inspection Evidence

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `8e8eb3ea5730681642b2406ca5d136ed8132ff59`
- inspected_files: `README.md`, `package.json`, `docs/archive/npm-rust-migration.md`, `docs/archive/rust-sqlite-vec-hooks.md`, `docs/bench-vault-fts.md`, `docs/claude-code-plugin.md`, `docs/development-workflow.md`, `docs/distribution/connect-in-2-minutes.md`, `docs/rust-distribution.md`, `packages/local/.mcp.json`, `packages/local/README.md`, `packages/local/assets/agent-rules.md`, `packages/local/assets/setup-prompt.md`, `packages/local/assets/skills/compile-context/skill.md`, `packages/local/assets/skills/vault-setup/skill.md`, `packages/local/bin/cli.js`, `packages/local/package.json`, `packages/sdk/README.md`, `packages/sdk/package.json`, `packages/sdk/src/hosted.ts`

Host AI hard rules:
- Without repo_clone_verified=true, do not claim that the source code has been read.
- Without repo_inspection_verified=true, do not write README, docs, or package-file conclusions as facts.
- Without quick_start_verified=true, do not claim that the Quick Start path has run successfully.

## Doramagic Pitfall Constraints

These rules come from Doramagic discovery, validation, or compilation findings. The host AI must treat them as operating constraints, not background notes.

### Constraint 1: Installation risk requires verification

- Trigger: Developers should check this installation risk before relying on the project: Explore FastMCP (Python) rewrite for MCP server
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: Explore FastMCP (Python) rewrite for MCP server. Context: Observed when using node, python
- Why it matters: Developers may fail before the first successful local run: Explore FastMCP (Python) rewrite for MCP server
- Evidence: failure_mode_cluster:github_issue | https://github.com/fellanH/context-vault/issues/191
- Hard boundary: Do not present this pitfall as solved, verified, or ignorable unless later evidence explicitly closes it.

### Constraint 2: Installation risk requires verification

- Trigger: Developers should check this installation risk before relying on the project: v3.12.0
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v3.12.0. Context: Observed when using node
- Why it matters: Upgrade or migration may change expected behavior: v3.12.0
- Evidence: failure_mode_cluster:github_release | https://github.com/fellanH/context-vault/releases/tag/v3.12.0
- Hard boundary: Do not present this pitfall as solved, verified, or ignorable unless later evidence explicitly closes it.

### Constraint 3: Installation risk requires verification

- Trigger: Developers should check this installation risk before relying on the project: v3.13.0
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v3.13.0. Context: Observed when using node
- Why it matters: Upgrade or migration may change expected behavior: v3.13.0
- Evidence: failure_mode_cluster:github_release | https://github.com/fellanH/context-vault/releases/tag/v3.13.0
- Hard boundary: Do not present this pitfall as solved, verified, or ignorable unless later evidence explicitly closes it.

### Constraint 4: Installation risk requires verification

- Trigger: Developers should check this installation risk before relying on the project: v3.16.1
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v3.16.1. Context: Observed during installation or first-run setup.
- Why it matters: Upgrade or migration may change expected behavior: v3.16.1
- Evidence: failure_mode_cluster:github_release | https://github.com/fellanH/context-vault/releases/tag/v3.16.1
- Hard boundary: Do not present this pitfall as solved, verified, or ignorable unless later evidence explicitly closes it.

### Constraint 5: Installation risk requires verification

- Trigger: Developers should check this installation risk before relying on the project: v3: Clean local/hosted separation — pure engine core
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v3: Clean local/hosted separation — pure engine core. Context: Observed when using node
- Why it matters: Developers may fail before the first successful local run: v3: Clean local/hosted separation — pure engine core
- Evidence: failure_mode_cluster:github_issue | https://github.com/fellanH/context-vault/issues/185
- Hard boundary: Do not present this pitfall as solved, verified, or ignorable unless later evidence explicitly closes it.

### Constraint 6: Installation risk requires verification

- Trigger: Developers should check this installation risk before relying on the project: v4.0.0
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v4.0.0. Context: Observed when using node, windows
- Why it matters: Upgrade or migration may change expected behavior: v4.0.0
- Evidence: failure_mode_cluster:github_release | https://github.com/fellanH/context-vault/releases/tag/v4.0.0
- Hard boundary: Do not present this pitfall as solved, verified, or ignorable unless later evidence explicitly closes it.