diff --git a/.githooks/post-checkout b/.githooks/post-checkout
new file mode 100755
index 00000000..573833dd
--- /dev/null
+++ b/.githooks/post-checkout
@@ -0,0 +1,8 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+repo_root="$({
+  cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd
+})"
+
+"$repo_root/scripts/ensure_agent_hardlinks.sh" >/dev/null
diff --git a/.githooks/post-merge b/.githooks/post-merge
new file mode 100755
index 00000000..573833dd
--- /dev/null
+++ b/.githooks/post-merge
@@ -0,0 +1,8 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+repo_root="$({
+  cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd
+})"
+
+"$repo_root/scripts/ensure_agent_hardlinks.sh" >/dev/null
diff --git a/.github/copilot-instructions.md b/.github/copilot-instructions.md
deleted file mode 120000
index be77ac83..00000000
--- a/.github/copilot-instructions.md
+++ /dev/null
@@ -1 +0,0 @@
-../AGENTS.md
\ No newline at end of file
diff --git a/.github/copilot-instructions.md b/.github/copilot-instructions.md
new file mode 100644
index 00000000..61cb6e0b
--- /dev/null
+++ b/.github/copilot-instructions.md
@@ -0,0 +1,216 @@
+# AGENTS.md
+
+> Canonical AI agent instructions for ColdVox. All tools (Claude Code, Copilot, Cursor, Kilo Code, etc.) should read this file.
+
+> **⚠️ CRITICAL**: Some documented features are broken or removed. See [`docs/plans/critical-action-plan.md`](docs/plans/critical-action-plan.md) before following STT instructions.
+
+> **Solo-dev trust model**: This repo is effectively single-developer. We optimize for local productivity/automation over “untrusted contributor” safety. If a tool/setup step enables automation (e.g., repo-managed git hooks), that’s intentional.
+
+## Key Docs
+
+- [`README.md`](README.md) - Project overview and quick start
+- [`CLAUDE.md`](CLAUDE.md) - Deep-dive guide (Claude-specific + imports AGENTS)
+- [`docs/plans/critical-action-plan.md`](docs/plans/critical-action-plan.md) - What’s currently broken/misleading
+- [`docs/dev/CI/architecture.md`](docs/dev/CI/architecture.md) - CI runner split and rationale
+- [`docs/standards.md`](docs/standards.md) - Documentation + changelog standards
+- [`docs/MasterDocumentationPlaybook.md`](docs/MasterDocumentationPlaybook.md) - Documentation structure and rules
+
+## Project Overview
+
+ColdVox is a Rust-based voice AI pipeline: audio capture → VAD → STT → text injection. Multi-crate Cargo workspace under `crates/`.
+
+**Key crates**: `coldvox-app` (main), `coldvox-audio`, `coldvox-vad`, `coldvox-vad-silero`, `coldvox-stt`, `coldvox-text-injection`, `coldvox-telemetry`, `coldvox-foundation`, `coldvox-gui`
+
+## Worktrees
+
+Use git worktrees for parallel agent work. This allows multiple agents to work on independent tasks simultaneously.
+
+```bash
+# Create worktree for a new task
+git worktree add ../.trees/coldvox-{task} -b {task}
+cd ../.trees/coldvox-{task}
+
+# List all worktrees
+git worktree list
+
+# Remove when done (after merge)
+git worktree remove ../.trees/coldvox-{task}
+```
+
+**Convention**: Worktrees live in `../.trees/coldvox-{branch-name}` to keep them adjacent but separate.
+
+## Commands
+
+### File-Scoped (Preferred)
+
+Always prefer file/crate-scoped commands over full workspace commands for faster feedback:
+
+```bash
+# Type check single crate
+cargo check -p coldvox-stt
+
+# Clippy single crate
+cargo clippy -p coldvox-audio 
+# Test single crate
+cargo test -p coldvox-text-injection
+
+# Format check (always full, it's fast)
+cargo fmt --all -- --check
+```
+
+### Full Workspace (When Needed)
+
+```bash
+just lint          # fmt + clippy + check (pre-push)
+just test          # cargo test --workspace --locked
+just build         # cargo build --workspace --locked
+just ci            # Full CI mirror via ./scripts/local_ci.sh
+```
+
+### Running
+
+```bash
+just run           # Main app
+just tui           # TUI dashboard
+cargo run --features text-injection  # Text injection enabled
+```
+
+## Do
+
+- Use `just lint` before every push
+- Prefer crate-scoped commands for faster iteration
+- Use feature flags: `whisper`, `parakeet`, `text-injection`, `silero`
+- Follow Conventional Commits: `feat:`, `fix:`, `docs:`, `refactor:`, `test:`
+- Run `cargo fmt --all` before committing
+- Add tests for new functionality
+- Update `CHANGELOG.md` for user-visible changes (see `docs/standards.md`)
+
+## Don't
+
+- Don't run full workspace builds when crate-scoped works
+- Don't commit secrets or `.env` files
+- Don't edit generated code under `target/`
+- Don't add heavy dependencies without discussion
+- Don't skip `just lint` before pushing
+- Don't create `docs/agents.md` - agent config lives at repo root
+
+## Project Structure
+
+```
+crates/
+  app/                    # Main application, binaries, integration
+  coldvox-audio/          # Audio capture, ring buffer, resampling
+  coldvox-vad/            # VAD traits and config
+  coldvox-vad-silero/     # Silero V5 ONNX VAD implementation
+  coldvox-stt/            # STT abstractions and plugins (Whisper, Parakeet)
+  coldvox-text-injection/ # Platform-specific text injection backends
+  coldvox-telemetry/      # Pipeline metrics
+  coldvox-foundation/     # Core types, error handling, shutdown
+  coldvox-gui/            # GUI components
+
+config/                   # Runtime configuration
+docs/                     # Documentation (see MasterDocumentationPlaybook.md)
+scripts/                  # Automation scripts
+```
+
+## Feature Flags
+
+Default: `silero`, `text-injection`
+
+- `whisper` - Faster-Whisper STT (Python-based, CPU/GPU)
+- `parakeet` - NVIDIA Parakeet STT (GPU-only, pure Rust)
+- `text-injection` - Platform-aware text injection
+- `silero` - Silero V5 ONNX VAD (default)
+- `examples` - Example binaries
+- `live-hardware-tests` - Hardware test suites
+
+## Safety & Permissions
+
+**Allowed without prompt:**
+- Read files, list files, search code
+- Crate-scoped: check, clippy, test, fmt
+- Git status, diff, log
+
+**Ask first:**
+- Package/dependency changes
+- Git push, force operations
+- Deleting files
+- Full workspace builds (prefer crate-scoped)
+- Database migrations
+- Running with hardware features
+
+## When Stuck
+
+- Ask a clarifying question
+- Propose a short plan before large changes
+- Open a draft PR with notes
+- Don't push large speculative changes without confirmation
+
+## Key Files
+
+- **Main entry**: `crates/app/src/main.rs`
+- **Audio pipeline**: `crates/coldvox-audio/src/capture.rs`
+- **VAD engine**: `crates/coldvox-vad-silero/src/silero_wrapper.rs`
+- **STT plugins**: `crates/coldvox-stt/src/plugins/`
+- **Text injection**: `crates/coldvox-text-injection/src/manager.rs`
+- **Build detection**: `crates/app/build.rs` (platform detection)
+
+## Documentation
+
+- `docs/MasterDocumentationPlaybook.md` - Documentation standards
+- `docs/standards.md` - Changelog rubric, metadata requirements
+- `docs/architecture.md` - System design and future vision
+- `docs/domains/` - Domain-specific technical docs
+
+## CI Environment
+
+> **Principle**: The self hosted runner on the laptop is used for hardware tests, CI for everything else.
+
+See [CI Architecture](docs/dev/CI/architecture.md) for full details.
+
+### Self-Hosted Runner (Fedora/Nobara)
+
+| Fact | Implication |
+|------|-------------|
+| **Live KDE Plasma session** | NO Xvfb needed. Use real `$DISPLAY`. |
+| **Fedora-based** | `apt-get` does NOT exist. Use `dnf`. |
+| **Always available** | Auto-login, survives reboots. |
+| **Warm sccache** | Incremental builds ~2-3 min. |
+
+### CI Split
+
+| Task | Runner | Why |
+|------|--------|-----|
+| `cargo fmt`, `cargo clippy` | GitHub-hosted | Fast, parallel, free |
+| `cargo audit`, `cargo deny` | GitHub-hosted | Security checks, no build needed |
+| `cargo build` | **Self-hosted** | Warm cache, THE build |
+| Hardware tests | **Self-hosted** | Requires display/audio/clipboard |
+
+### DON'T (Common AI Mistakes)
+
+```yaml
+# WRONG - runner has live display, also uses apt-get internally
+- uses: GabrielBB/xvfb-action@v1
+
+# WRONG - this is Fedora, not Ubuntu
+- run: sudo apt-get install -y xdotool
+
+# WRONG - real display is :0
+env:
+  DISPLAY: ":99"
+
+# WRONG - delays self-hosted by 5-10 min
+hardware:
+  needs: [lint, build]
+
+# WRONG - wasted work, can't share artifacts with Fedora
+- run: cargo build  # On ubuntu-latest
+```
+
+## PR Checklist
+
+- [ ] `just lint` passes
+- [ ] `just test` passes (or crate-scoped tests)
+- [ ] Changelog updated if user-visible
+- [ ] Commit messages follow Conventional Commits
+- [ ] No secrets or sensitive data
diff --git a/.github/workflows/docs-ci.yml b/.github/workflows/docs-ci.yml
index 89786a64..3a407d50 100644
--- a/.github/workflows/docs-ci.yml
+++ b/.github/workflows/docs-ci.yml
@@ -39,6 +39,11 @@ jobs:
         run: |
           uv run scripts/validate_docs.py "${{ github.event.pull_request.base.sha }}" "${{ github.sha }}"
 
+      - name: Validate agent instruction mirrors
+        run: |
+          diff -u AGENTS.md .github/copilot-instructions.md
+          diff -u AGENTS.md .kilocode/rules/agents.md
+
       - name: Append revision log entries
         run: |
           uv run scripts/docs_revision_logger.py "${{ github.event.pull_request.base.sha }}" "${{ github.sha }}" "${{ github.actor }}"
diff --git a/.kilocode/rules/agents.md b/.kilocode/rules/agents.md
deleted file mode 120000
index b7e6491d..00000000
--- a/.kilocode/rules/agents.md
+++ /dev/null
@@ -1 +0,0 @@
-../../AGENTS.md
\ No newline at end of file
diff --git a/.kilocode/rules/agents.md b/.kilocode/rules/agents.md
new file mode 100644
index 00000000..61cb6e0b
--- /dev/null
+++ b/.kilocode/rules/agents.md
@@ -0,0 +1,216 @@
+# AGENTS.md
+
+> Canonical AI agent instructions for ColdVox. All tools (Claude Code, Copilot, Cursor, Kilo Code, etc.) should read this file.
+
+> **⚠️ CRITICAL**: Some documented features are broken or removed. See [`docs/plans/critical-action-plan.md`](docs/plans/critical-action-plan.md) before following STT instructions.
+
+> **Solo-dev trust model**: This repo is effectively single-developer. We optimize for local productivity/automation over “untrusted contributor” safety. If a tool/setup step enables automation (e.g., repo-managed git hooks), that’s intentional.
+
+## Key Docs
+
+- [`README.md`](README.md) - Project overview and quick start
+- [`CLAUDE.md`](CLAUDE.md) - Deep-dive guide (Claude-specific + imports AGENTS)
+- [`docs/plans/critical-action-plan.md`](docs/plans/critical-action-plan.md) - What’s currently broken/misleading
+- [`docs/dev/CI/architecture.md`](docs/dev/CI/architecture.md) - CI runner split and rationale
+- [`docs/standards.md`](docs/standards.md) - Documentation + changelog standards
+- [`docs/MasterDocumentationPlaybook.md`](docs/MasterDocumentationPlaybook.md) - Documentation structure and rules
+
+## Project Overview
+
+ColdVox is a Rust-based voice AI pipeline: audio capture → VAD → STT → text injection. Multi-crate Cargo workspace under `crates/`.
+
+**Key crates**: `coldvox-app` (main), `coldvox-audio`, `coldvox-vad`, `coldvox-vad-silero`, `coldvox-stt`, `coldvox-text-injection`, `coldvox-telemetry`, `coldvox-foundation`, `coldvox-gui`
+
+## Worktrees
+
+Use git worktrees for parallel agent work. This allows multiple agents to work on independent tasks simultaneously.
+
+```bash
+# Create worktree for a new task
+git worktree add ../.trees/coldvox-{task} -b {task}
+cd ../.trees/coldvox-{task}
+
+# List all worktrees
+git worktree list
+
+# Remove when done (after merge)
+git worktree remove ../.trees/coldvox-{task}
+```
+
+**Convention**: Worktrees live in `../.trees/coldvox-{branch-name}` to keep them adjacent but separate.
+
+## Commands
+
+### File-Scoped (Preferred)
+
+Always prefer file/crate-scoped commands over full workspace commands for faster feedback:
+
+```bash
+# Type check single crate
+cargo check -p coldvox-stt
+
+# Clippy single crate
+cargo clippy -p coldvox-audio 
+# Test single crate
+cargo test -p coldvox-text-injection
+
+# Format check (always full, it's fast)
+cargo fmt --all -- --check
+```
+
+### Full Workspace (When Needed)
+
+```bash
+just lint          # fmt + clippy + check (pre-push)
+just test          # cargo test --workspace --locked
+just build         # cargo build --workspace --locked
+just ci            # Full CI mirror via ./scripts/local_ci.sh
+```
+
+### Running
+
+```bash
+just run           # Main app
+just tui           # TUI dashboard
+cargo run --features text-injection  # Text injection enabled
+```
+
+## Do
+
+- Use `just lint` before every push
+- Prefer crate-scoped commands for faster iteration
+- Use feature flags: `whisper`, `parakeet`, `text-injection`, `silero`
+- Follow Conventional Commits: `feat:`, `fix:`, `docs:`, `refactor:`, `test:`
+- Run `cargo fmt --all` before committing
+- Add tests for new functionality
+- Update `CHANGELOG.md` for user-visible changes (see `docs/standards.md`)
+
+## Don't
+
+- Don't run full workspace builds when crate-scoped works
+- Don't commit secrets or `.env` files
+- Don't edit generated code under `target/`
+- Don't add heavy dependencies without discussion
+- Don't skip `just lint` before pushing
+- Don't create `docs/agents.md` - agent config lives at repo root
+
+## Project Structure
+
+```
+crates/
+  app/                    # Main application, binaries, integration
+  coldvox-audio/          # Audio capture, ring buffer, resampling
+  coldvox-vad/            # VAD traits and config
+  coldvox-vad-silero/     # Silero V5 ONNX VAD implementation
+  coldvox-stt/            # STT abstractions and plugins (Whisper, Parakeet)
+  coldvox-text-injection/ # Platform-specific text injection backends
+  coldvox-telemetry/      # Pipeline metrics
+  coldvox-foundation/     # Core types, error handling, shutdown
+  coldvox-gui/            # GUI components
+
+config/                   # Runtime configuration
+docs/                     # Documentation (see MasterDocumentationPlaybook.md)
+scripts/                  # Automation scripts
+```
+
+## Feature Flags
+
+Default: `silero`, `text-injection`
+
+- `whisper` - Faster-Whisper STT (Python-based, CPU/GPU)
+- `parakeet` - NVIDIA Parakeet STT (GPU-only, pure Rust)
+- `text-injection` - Platform-aware text injection
+- `silero` - Silero V5 ONNX VAD (default)
+- `examples` - Example binaries
+- `live-hardware-tests` - Hardware test suites
+
+## Safety & Permissions
+
+**Allowed without prompt:**
+- Read files, list files, search code
+- Crate-scoped: check, clippy, test, fmt
+- Git status, diff, log
+
+**Ask first:**
+- Package/dependency changes
+- Git push, force operations
+- Deleting files
+- Full workspace builds (prefer crate-scoped)
+- Database migrations
+- Running with hardware features
+
+## When Stuck
+
+- Ask a clarifying question
+- Propose a short plan before large changes
+- Open a draft PR with notes
+- Don't push large speculative changes without confirmation
+
+## Key Files
+
+- **Main entry**: `crates/app/src/main.rs`
+- **Audio pipeline**: `crates/coldvox-audio/src/capture.rs`
+- **VAD engine**: `crates/coldvox-vad-silero/src/silero_wrapper.rs`
+- **STT plugins**: `crates/coldvox-stt/src/plugins/`
+- **Text injection**: `crates/coldvox-text-injection/src/manager.rs`
+- **Build detection**: `crates/app/build.rs` (platform detection)
+
+## Documentation
+
+- `docs/MasterDocumentationPlaybook.md` - Documentation standards
+- `docs/standards.md` - Changelog rubric, metadata requirements
+- `docs/architecture.md` - System design and future vision
+- `docs/domains/` - Domain-specific technical docs
+
+## CI Environment
+
+> **Principle**: The self hosted runner on the laptop is used for hardware tests, CI for everything else.
+
+See [CI Architecture](docs/dev/CI/architecture.md) for full details.
+
+### Self-Hosted Runner (Fedora/Nobara)
+
+| Fact | Implication |
+|------|-------------|
+| **Live KDE Plasma session** | NO Xvfb needed. Use real `$DISPLAY`. |
+| **Fedora-based** | `apt-get` does NOT exist. Use `dnf`. |
+| **Always available** | Auto-login, survives reboots. |
+| **Warm sccache** | Incremental builds ~2-3 min. |
+
+### CI Split
+
+| Task | Runner | Why |
+|------|--------|-----|
+| `cargo fmt`, `cargo clippy` | GitHub-hosted | Fast, parallel, free |
+| `cargo audit`, `cargo deny` | GitHub-hosted | Security checks, no build needed |
+| `cargo build` | **Self-hosted** | Warm cache, THE build |
+| Hardware tests | **Self-hosted** | Requires display/audio/clipboard |
+
+### DON'T (Common AI Mistakes)
+
+```yaml
+# WRONG - runner has live display, also uses apt-get internally
+- uses: GabrielBB/xvfb-action@v1
+
+# WRONG - this is Fedora, not Ubuntu
+- run: sudo apt-get install -y xdotool
+
+# WRONG - real display is :0
+env:
+  DISPLAY: ":99"
+
+# WRONG - delays self-hosted by 5-10 min
+hardware:
+  needs: [lint, build]
+
+# WRONG - wasted work, can't share artifacts with Fedora
+- run: cargo build  # On ubuntu-latest
+```
+
+## PR Checklist
+
+- [ ] `just lint` passes
+- [ ] `just test` passes (or crate-scoped tests)
+- [ ] Changelog updated if user-visible
+- [ ] Commit messages follow Conventional Commits
+- [ ] No secrets or sensitive data
diff --git a/AGENTS.md b/AGENTS.md
index bb1b989f..61cb6e0b 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -2,7 +2,18 @@
 
 > Canonical AI agent instructions for ColdVox. All tools (Claude Code, Copilot, Cursor, Kilo Code, etc.) should read this file.
 
-> **⚠️ CRITICAL**: Some documented features are broken or removed. See [`criticalActionPlan.md`](criticalActionPlan.md) before following STT instructions.
+> **⚠️ CRITICAL**: Some documented features are broken or removed. See [`docs/plans/critical-action-plan.md`](docs/plans/critical-action-plan.md) before following STT instructions.
+
+> **Solo-dev trust model**: This repo is effectively single-developer. We optimize for local productivity/automation over “untrusted contributor” safety. If a tool/setup step enables automation (e.g., repo-managed git hooks), that’s intentional.
+
+## Key Docs
+
+- [`README.md`](README.md) - Project overview and quick start
+- [`CLAUDE.md`](CLAUDE.md) - Deep-dive guide (Claude-specific + imports AGENTS)
+- [`docs/plans/critical-action-plan.md`](docs/plans/critical-action-plan.md) - What’s currently broken/misleading
+- [`docs/dev/CI/architecture.md`](docs/dev/CI/architecture.md) - CI runner split and rationale
+- [`docs/standards.md`](docs/standards.md) - Documentation + changelog standards
+- [`docs/MasterDocumentationPlaybook.md`](docs/MasterDocumentationPlaybook.md) - Documentation structure and rules
 
 ## Project Overview
 
@@ -61,7 +72,7 @@ just ci            # Full CI mirror via ./scripts/local_ci.sh
 ```bash
 just run           # Main app
 just tui           # TUI dashboard
-cargo run --features whisper,text-injection  # With STT
+cargo run --features text-injection  # Text injection enabled
 ```
 
 ## Do
diff --git a/CLAUDE.md b/CLAUDE.md
index 0ffc5dcc..bb73ee4e 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -2,7 +2,7 @@
 
 This file provides guidance to Claude Code (claude.ai/code) when working with this repository.
 
-> **⚠️ CRITICAL**: STT documentation below is outdated. Whisper is removed, Parakeet broken. See [`criticalActionPlan.md`](criticalActionPlan.md). **Only Moonshine works.**
+> **⚠️ CRITICAL**: STT documentation below is outdated. Whisper is removed, Parakeet broken. See [`docs/plans/critical-action-plan.md`](docs/plans/critical-action-plan.md). **Only Moonshine works.**
 
 **@import AGENTS.md** - Read `AGENTS.md` for canonical project instructions (structure, commands, do/don't rules, worktrees).
 
diff --git a/README.md b/README.md
index ffc65db9..a20604e1 100644
--- a/README.md
+++ b/README.md
@@ -1,7 +1,7 @@
 # ColdVox
 > ⚠️ **Internal Alpha** - This project is in early development and not ready for production use.
 
-> **⚠️ CRITICAL**: Documentation is out of sync with code. Whisper STT has been removed; Parakeet doesn't compile. See [`criticalActionPlan.md`](criticalActionPlan.md) for current status. **Only Moonshine STT works** (requires `uv sync` first).
+> **⚠️ CRITICAL**: Documentation is out of sync with code. Whisper STT has been removed; Parakeet doesn't compile. See [`docs/plans/critical-action-plan.md`](docs/plans/critical-action-plan.md) for current status. **Only Moonshine STT works** (requires `uv sync` first).
 ## Development
  - Install Rust (stable) and required system dependencies for your platform.
  - Use the provided scripts in `scripts/` to help with local environment setup.
@@ -113,4 +113,4 @@ Dual-licensed under MIT or Apache-2.0. See `LICENSE-MIT` and `LICENSE-APACHE` if
 - Review the [Master Documentation Playbook](docs/MasterDocumentationPlaybook.md).
 - Follow the repository [Documentation Standards](docs/standards.md).
 - Coordinate work through the [Documentation Todo Backlog](docs/todo.md).
-- Assistants should read the [Assistant Interaction Index](docs/agents.md).
+- Assistants should read [`AGENTS.md`](AGENTS.md).
diff --git a/docs/architecture.md b/docs/architecture.md
index 6a79f848..e5a656f9 100644
--- a/docs/architecture.md
+++ b/docs/architecture.md
@@ -9,13 +9,13 @@ last_reviewed: 2025-10-19
 
 # ColdVox Architecture & Future Vision
 
-> **⚠️ CRITICAL**: STT backend status has changed. See [`../criticalActionPlan.md`](../criticalActionPlan.md) for current working features.
+> **⚠️ CRITICAL**: STT backend status has changed. See [`plans/critical-action-plan.md`](plans/critical-action-plan.md) for current working features.
 
 ## Navigation
 
 - [Architecture Roadmap](./architecture/roadmap.md)
 - [Architecture Decisions](./architecture/adr/)
-- [Critical Action Plan](../criticalActionPlan.md) - Current broken features tracking
+- [Critical Action Plan](plans/critical-action-plan.md) - Current broken features tracking
 
 
 
diff --git a/docs/architecture/roadmap.md b/docs/architecture/roadmap.md
index 7c3f57b7..b6fa24b5 100644
--- a/docs/architecture/roadmap.md
+++ b/docs/architecture/roadmap.md
@@ -53,7 +53,7 @@ This roadmap outlines the anticipated progression of major transformational init
 
 ## Next Actions (Short-Term)
 
-1. Finalize `docs/standards.md`, `docs/agents.md`, and playbook skeletons.
+1. Finalize `docs/standards.md`, `AGENTS.md`, and playbook skeletons.
 2. Land GitHub governance automation plans in `docs/playbooks/organizational/github_governance.md`.
 3. Prototype docs watcher tooling and integrate logs into CI for visibility.
 4. Draft the always-on listening proof-of-concept design notes and spike tickets for v1.0 foundation work.
diff --git a/criticalActionPlan.md b/docs/plans/critical-action-plan.md
similarity index 100%
rename from criticalActionPlan.md
rename to docs/plans/critical-action-plan.md
diff --git a/docs/plans/documentation/future-documentation-architecture.md b/docs/plans/documentation/future-documentation-architecture.md
index 67c6ca15..8cbd387a 100644
--- a/docs/plans/documentation/future-documentation-architecture.md
+++ b/docs/plans/documentation/future-documentation-architecture.md
@@ -87,7 +87,7 @@ docs/
 - **Audience**: All contributors
 - **Enforcement**: CI validation
 
-#### 3. Agent Guidelines (`docs/agents.md`)
+#### 3. Agent Guidelines (`AGENTS.md`)
 - **Purpose**: Optimize documentation for AI assistants
 - **Content**: Documentation index, search patterns, contribution guidelines
 - **Audience**: AI agents (Claude, GitHub Copilot, etc.)
@@ -166,7 +166,7 @@ last_reviewed: [YYYY-MM-DD]
 
 ### For AI Agents
 - **Predictable Structure**: Domain-based organization matches code structure
-- **Comprehensive Index**: `docs/agents.md` provides navigation guidance
+- **Comprehensive Index**: `AGENTS.md` provides navigation guidance
 - **Standardized Format**: Consistent headers and metadata enable better parsing
 - **Process Guidance**: Playbooks provide context for contribution workflows
 
diff --git a/agentsDocResearch.md b/docs/research/logs/2025-12-25-agents-doc-research.md
similarity index 100%
rename from agentsDocResearch.md
rename to docs/research/logs/2025-12-25-agents-doc-research.md
diff --git a/toolEditResearch.md b/docs/research/logs/2025-12-25-change-scoped-file-editing.md
similarity index 100%
rename from toolEditResearch.md
rename to docs/research/logs/2025-12-25-change-scoped-file-editing.md
diff --git a/PR-190-Comprehensive-Assessment.md b/docs/research/pr-reports/PR-190-Comprehensive-Assessment.md
similarity index 100%
rename from PR-190-Comprehensive-Assessment.md
rename to docs/research/pr-reports/PR-190-Comprehensive-Assessment.md
diff --git a/justfile b/justfile
index 20370f4e..d7360468 100644
--- a/justfile
+++ b/justfile
@@ -66,6 +66,14 @@ docs-validate base="origin/main" head="HEAD":
 setup-hooks:
     pre-commit install
 
+# Ensure agent instruction mirrors are hardlinked to AGENTS.md
+link-agents:
+    ./scripts/ensure_agent_hardlinks.sh
+
+# Install git hooks to keep agent hardlinks after checkout/merge
+setup-hardlink-hooks:
+    ./scripts/install_git_hardlink_hooks.sh
+
 # Skip Rust checks in pre-commit (useful for quick commits)
 commit-fast *args:
     SKIP_RUST_CHECKS=1 git commit {{args}}
@@ -105,7 +113,7 @@ setup-sccache:
     echo "To enable: export RUSTC_WRAPPER=sccache"
 
 # Setup all development tools (run once after clone)
-setup: setup-hooks setup-sccache
+setup: setup-hooks setup-sccache setup-hardlink-hooks link-agents
     @echo "✓ Development environment ready"
 
 # Install Moonshine Python dependencies (transformers, torch, librosa via uv)
diff --git a/mise.toml b/mise.toml
index 886e9c09..ca02293a 100644
--- a/mise.toml
+++ b/mise.toml
@@ -33,6 +33,19 @@ description = "Universal Git Pre-commit Hook"
 # Avoids 'npx' to prevent registry lookup latency.
 run = "lint-staged"
 
+[tasks.prepare]
+description = "Repo bootstrap: hooks + agent hardlinks"
+run = """
+#!/usr/bin/env bash
+set -euo pipefail
+
+# Ensure pre-commit hooks are installed (idempotent).
+pre-commit install
+
+# Enable repo-tracked git hooks and ensure AGENTS mirrors are hardlinked.
+./scripts/install_git_hardlink_hooks.sh >/dev/null
+"""
+
 # --- TASK DELEGATION LAYER ---
 # lint-staged delegates logic here.
 
diff --git a/scripts/ensure_agent_hardlinks.sh b/scripts/ensure_agent_hardlinks.sh
new file mode 100755
index 00000000..49bfe1e2
--- /dev/null
+++ b/scripts/ensure_agent_hardlinks.sh
@@ -0,0 +1,26 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+repo_root="$({
+  cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd
+})"
+
+src="$repo_root/AGENTS.md"
+
+dst1="$repo_root/.github/copilot-instructions.md"
+dst2="$repo_root/.kilocode/rules/agents.md"
+
+if [[ ! -f "$src" ]]; then
+  echo "error: missing $src" >&2
+  exit 1
+fi
+
+mkdir -p "$(dirname "$dst1")" "$(dirname "$dst2")"
+
+# Replace destination files with hard links to AGENTS.md.
+# This is safe even if Git doesn't preserve hardlinks across clones: rerun anytime.
+ln -f "$src" "$dst1"
+ln -f "$src" "$dst2"
+
+# Show inode + link count for confirmation.
+stat -c '%i %h %n' "$src" "$dst1" "$dst2"
diff --git a/scripts/install_git_hardlink_hooks.sh b/scripts/install_git_hardlink_hooks.sh
new file mode 100755
index 00000000..342d361a
--- /dev/null
+++ b/scripts/install_git_hardlink_hooks.sh
@@ -0,0 +1,25 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+repo_root="$({
+  cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd
+})"
+
+if [[ ! -d "$repo_root/.git" ]]; then
+  echo "error: not a git checkout (missing $repo_root/.git)" >&2
+  exit 1
+fi
+
+hooks_path="$repo_root/.githooks"
+
+if [[ ! -d "$hooks_path" ]]; then
+  echo "error: missing $hooks_path (expected repo-tracked hooks)" >&2
+  exit 1
+fi
+
+chmod +x "$hooks_path/post-checkout" "$hooks_path/post-merge" "$repo_root/scripts/ensure_agent_hardlinks.sh"
+
+git -C "$repo_root" config core.hooksPath .githooks
+
+echo "Enabled repo hooks via core.hooksPath=.githooks"
+"$repo_root/scripts/ensure_agent_hardlinks.sh"