### Install Roborev with Go

Source: https://github.com/roborev-dev/roborev/blob/main/README.md

Install the Roborev CLI tool using the Go build tools.

```bash
go install github.com/roborev-dev/roborev/cmd/roborev@latest
```

--------------------------------

### Initialize Test Repository and Get HEAD SHA

Source: https://context7.com/roborev-dev/roborev/llms.txt

Standard setup for tests that initializes a Git repository with a base commit and retrieves the SHA of the HEAD commit.

```go
dir := testutil.InitTestRepo(t)
sha := testutil.GetHeadSHA(t, dir)
```

--------------------------------

### Global configuration example

Source: https://context7.com/roborev-dev/roborev/llms.txt

Example of a `~/.roborev/config.toml` file for global Roborev settings. This includes the default agent, model, API keys, and worker count.

```toml
# ~/.roborev/config.toml

# Default agent (used when no per-repo config is present)
default_agent = "claude-code"

# Default model
default_model = ""

# Anthropic API key (used by claude-code; re-injected into agent environment)
anthropic_api_key = "sk-ant-..."

# Maximum prompt size across all repos (bytes)
default_max_prompt_size = 204800

# Number of parallel workers
workers = 4

# Global review guidelines (appended to every review)
review_guidelines = ""

# Allow unsafe agents globally
allow_unsafe_agents = false
```

--------------------------------

### Initialize roborev repository

Source: https://context7.com/roborev-dev/roborev/llms.txt

Initialize roborev in a git repository. This sets up global and local configuration, installs git hooks, and starts the background daemon.

```bash
# Basic initialization (auto-detect agent from PATH)
cd your-repo
roborev init
```

```bash
# Initialize with a specific agent written to .roborev.toml
roborev init --agent claude-code
```

```bash
# Initialize without auto-starting the daemon (useful with systemd/launchd)
roborev init --no-daemon
```

```bash
# Expected output:
# Initializing roborev...
#   Created config at /home/user/.roborev/config.toml
#   Created .roborev.toml
#   Daemon is running
#   Repo registered
#
# Ready! Every commit will now be automatically reviewed.
```

--------------------------------

### Install roborev

Source: https://context7.com/roborev-dev/roborev/llms.txt

Install roborev using various methods including a shell script, Homebrew, Go install, or PowerShell.

```bash
# macOS / Linux (shell script)
curl -fsSL https://roborev.io/install.sh | bash
```

```bash
# Homebrew
brew install roborev-dev/tap/roborev
```

```bash
# Go install
go install github.com/roborev-dev/roborev/cmd/roborev@latest
```

```powershell
# Windows (PowerShell)
powershell -ExecutionPolicy ByPass -c "irm https://roborev.io/install.ps1 | iex"
```

--------------------------------

### Per-repo configuration example

Source: https://context7.com/roborev-dev/roborev/llms.txt

Example of a `.roborev.toml` file used for per-repository configuration overrides. It allows setting default agents, models, reasoning levels, and CI settings.

```toml
# .roborev.toml

# Default agent for all workflows
agent = "claude-code"

# Default model
model = "claude-sonnet-4-5"

# Fallback agent when primary hits quota
backup_agent = "gemini"

# Workflow-specific agent overrides
review_agent       = "claude-code"
fix_agent          = "codex"
refine_agent       = "claude-code"
security_agent     = "claude-code"
design_agent       = "claude-code"

# Workflow-specific reasoning levels
review_reasoning   = "thorough"    # default
fix_reasoning      = "standard"    # default
refine_reasoning   = "standard"    # default

# Model routing by reasoning tier
review_model_fast     = "claude-haiku-4-5"
review_model_thorough = "claude-opus-4-5"

# Route claude-code to a local Ollama proxy for reviews
review_model = "llama3:latest@http://127.0.0.1:11434"

# Project-specific review instructions appended to every review prompt
review_guidelines = """
Always check for missing error handling in public functions.
Enforce the project's logging conventions (structured JSON only).
"""

# Number of recent reviews to include as context
review_context_count = 3

# Per-job timeout in minutes (default: 30)
job_timeout_minutes = 45

# Maximum prompt size in bytes (default: 200KB = 204800)
max_prompt_size = 204800

# Minimum severity for review findings: critical, high, medium, low
review_min_severity = "medium"

# Post-commit review mode: "commit" (default) or "branch"
post_commit_review = "commit"

# Patterns to exclude from diffs (supports glob syntax)
exclude_patterns = ["*_generated.go", "vendor/**", "*.pb.go"]

# Allow agents that require file-write access (needed for refine)
allow_unsafe_agents = true

# CI integration
[ci]
enabled       = true
github_repo   = "myorg/myrepo"
poll_interval = "60s"
agents        = ["claude-code", "codex"]
review_types  = ["", "security"]
min_severity  = "high"

# Hooks: run shell commands on review events
[hooks]
on_review_failed  = "notify-send 'roborev' 'Review failed: {{.Subject}}'"
on_review_passed  = ""

```

--------------------------------

### Install agent skills

Source: https://context7.com/roborev-dev/roborev/llms.txt

Use `roborev skills install` to install skill files for agents, enabling them to invoke Roborev workflows. `roborev skills list` shows installed skills.

```bash
# Install skills for all detected agents
roborev skills install

# Install for a specific agent
roborev skills install --agent claude-code

# List installed skills
roborev skills list

```

--------------------------------

### Install and Run Pre-commit Hooks

Source: https://github.com/roborev-dev/roborev/blob/main/README.md

Install the prek tool and its git hook for local pre-commit checks. Run checks manually to ensure code quality before committing.

```bash
brew install prek     # or use your preferred prek install method
prek install          # install the local git hook
prek run --all-files  # run the configured checks manually
```

--------------------------------

### Install roborev Skills

Source: https://github.com/roborev-dev/roborev/blob/main/skills/README.md

Run this command to install the roborev skills. Skills are automatically updated when you run `roborev update`.

```bash
roborev skills install
```

--------------------------------

### Example: Design Review Against Specific Base

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/codex/roborev-design-review-branch/SKILL.md

Demonstrates the process for initiating a design review against a specified base branch, including validation and execution.

```bash
$roborev-design-review-branch --base develop
```

--------------------------------

### Roborev Configuration Example

Source: https://github.com/roborev-dev/roborev/blob/main/README.md

Configure Roborev by creating a .roborev.toml file. Specify the agent to use and provide project-specific review guidelines.

```toml
agent = "claude-code"
review_guidelines = """
Project-specific review instructions here.
"""
```

--------------------------------

### roborev skills install

Source: https://context7.com/roborev-dev/roborev/llms.txt

Installs agent skills into the configuration directory, allowing agents to invoke Roborev workflows directly. Includes commands to list installed skills.

```APIDOC
## `roborev skills install` — Install agent skills

Installs roborev skill files into the agent's configuration directory (Claude Code `~/.claude/` or Codex `~/.codex/`) so that agents can invoke roborev workflows directly from interactive sessions.

```bash
# Install skills for all detected agents
roborev skills install

# Install for a specific agent
roborev skills install --agent claude-code

# List installed skills
roborev skills list

# Skills installed (examples):
#   roborev:review         Queue a review from inside an agent session
#   roborev:fix            Discover and apply open review findings
#   roborev:refine         Run the full review-fix-review loop
#   roborev:design-review  Run a design-focused review
```
```

--------------------------------

### Install Roborev via Homebrew (macOS/Linux)

Source: https://github.com/roborev-dev/roborev/blob/main/README.md

Install Roborev using the Homebrew package manager on macOS or Linux.

```bash
brew install roborev-dev/tap/roborev
```

--------------------------------

### Security Review with Specific Base Example

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/codex/roborev-review-branch/SKILL.md

Example of invoking the skill for a security review against a specific base branch. Includes validation of the base ref and execution of the review command.

```bash
git rev-parse --verify -- develop
```

```bash
roborev review --branch --wait --base develop --type security
```

--------------------------------

### Example: Default Branch Review Command

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-review-branch/SKILL.md

Illustrates the command executed for a default branch review without specific options.

```bash
roborev review --branch --wait
```

--------------------------------

### Example: Respond to Review with Message

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-respond/SKILL.md

Demonstrates the agent action when a user provides a job ID and a message for the roborev-respond skill.

```bash
roborev comment --job 1019 "Fixed all issues" && roborev close 1019
```

--------------------------------

### Initialize Roborev Post-Commit Hook

Source: https://github.com/roborev-dev/roborev/blob/main/README.md

Run this command to install the git post-commit hook. Reviews are triggered automatically on every commit.

```bash
cd your-repo
roborev init          # Install post-commit hook
git commit -m "..."   # Reviews happen automatically
roborev tui           # View reviews in interactive UI
```

--------------------------------

### Run Go Integration Tests

Source: https://context7.com/roborev-dev/roborev/llms.txt

Command to execute integration tests, typically requiring more setup or external dependencies.

```bash
go test -tags=integration ./...
```

--------------------------------

### Example: Reviewing a Specific Design Document

Source: https://github.com/roborev-dev/roborev/blob/main/skills/roborev-design-review.md

This example demonstrates how to use the roborev-design-review skill to analyze a specific design document located at 'docs/design/auth-redesign.md'. The skill will then read the document, verify assumptions against source files, and review both the PRD and task list for various quality attributes.

```bash
/roborev-design-review docs/design/auth-redesign.md
```

--------------------------------

### Install Roborev via PowerShell (Windows)

Source: https://github.com/roborev-dev/roborev/blob/main/README.md

Use this PowerShell command to install Roborev on Windows systems.

```powershell
powershell -ExecutionPolicy ByPass -c "irm https://roborev.io/install.ps1 | iex"
```

--------------------------------

### Example: Security Review Against Specific Base Command

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-review-branch/SKILL.md

Demonstrates the command for a security review targeting a specific base branch.

```bash
roborev review --branch --wait --base develop --type security
```

--------------------------------

### Example: Default Branch Design Review

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/codex/roborev-design-review-branch/SKILL.md

Illustrates the user interaction and agent actions for a default branch design review, including presenting findings and offering to fix them.

```bash
$roborev-design-review-branch
```

--------------------------------

### Install Roborev via Shell Script (macOS/Linux)

Source: https://github.com/roborev-dev/roborev/blob/main/README.md

Use this command to install Roborev directly on macOS or Linux systems.

```bash
curl -fsSL https://roborev.io/install.sh | bash
```

--------------------------------

### Refine from a Specific Starting Commit

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-refine/SKILL.md

Initiates the roborev refine process starting from a specified commit. Allows setting a maximum number of iterations for the refinement process.

```bash
/roborev-refine --since abc123 --max-iterations 3
```

--------------------------------

### Example: Security Review of Specific Commit

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-review/SKILL.md

Illustrates the process for a security review of a specific commit, including validation, background task execution, and user notification.

```bash
roborev review abc123 --wait --type security
```

--------------------------------

### Example: Default Review of HEAD

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-review/SKILL.md

Demonstrates the agent's actions for a default review of the HEAD commit, including submitting the review, notifying the user, and offering to fix findings.

```bash
roborev review --wait
```

--------------------------------

### List registered repositories

Source: https://context7.com/roborev-dev/roborev/llms.txt

Get a list of all repositories currently registered with Roborev. The output is processed with `jq` for readability.

```shell
curl http://127.0.0.1:7373/api/repos | jq .
```

--------------------------------

### Example: Respond to Review without Message

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-respond/SKILL.md

Illustrates the agent action after prompting the user for a message when only a job ID is provided to the roborev-respond skill.

```bash
roborev comment --job 1019 "The null check was a false positive" && roborev close 1019
```

--------------------------------

### Get Activity Log

Source: https://github.com/roborev-dev/roborev/blob/main/CLAUDE.md

Retrieves the activity log of the daemon.

```APIDOC
## GET /activity

### Description
Activity log

### Method
GET

### Endpoint
/activity
```

--------------------------------

### Security Review of Specific Commit

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/codex/roborev-review/SKILL.md

Example of a security review for a specific commit. The skill first validates the commit reference, then executes the review with the specified type.

```bash
git rev-parse --verify -- abc123^{commit}
```

```bash
roborev review abc123 --wait --type security
```

--------------------------------

### Configure PostgreSQL Sync in TOML

Source: https://context7.com/roborev-dev/roborev/llms.txt

Configures the optional PostgreSQL synchronization by providing the database connection URL and the sync interval. This enables multi-machine setups using a shared database.

```toml
# ~/.roborev/config.toml
[sync]
postgres_url = "postgres://user:pass@db.example.com:5432/roborev"
sync_interval = "30s"
```

--------------------------------

### Daemon lifecycle management

Source: https://context7.com/roborev-dev/roborev/llms.txt

Manage the Roborev daemon using `roborev daemon` commands: `run` (foreground), `start` (background), `stop`, and `info`. The daemon manages workers, storage, and CI polling.

```bash
# Start the daemon in the foreground
roborev daemon run

# Start in the background (detached)
roborev daemon start

# Stop the background daemon
roborev daemon stop

# Show daemon runtime info (PID, address, port)
roborev daemon info

```

--------------------------------

### Get Available Agent with Configuration in Go

Source: https://context7.com/roborev-dev/roborev/llms.txt

Retrieves the first available agent from a fallback chain, respecting provided configurations. Allows chaining configuration methods like WithReasoning, WithAgentic, and WithModel.

```go
// Get first available agent from fallback chain (codex → claude-code → gemini → ...):
a, err := agent.GetAvailableWithConfig(repoPath, "claude-code", cfg, "codex")
a = a.WithReasoning(agent.ReasoningThorough).WithAgentic(true).WithModel("claude-opus-4-5")
```

--------------------------------

### HTTP API: Get review output

Source: https://context7.com/roborev-dev/roborev/llms.txt

Fetch the review output for a specific job using the `/api/review` endpoint with the `job_id` query parameter. The output is returned as a string.

```bash
curl "http://127.0.0.1:7373/api/review?job_id=42" | jq .output

```

--------------------------------

### roborev daemon

Source: https://context7.com/roborev-dev/roborev/llms.txt

Manages the lifecycle of the Roborev daemon, which includes an HTTP server, worker pool, and storage. Commands allow starting, stopping, and checking daemon status.

```APIDOC
## `roborev daemon` — Daemon lifecycle management

The daemon is an HTTP server that manages a worker pool, SQLite storage, CI polling, and SSE event streams. Most CLI commands auto-start it; these commands give explicit control.

```bash
# Start the daemon in the foreground
roborev daemon run

# Start in the background (detached)
roborev daemon start

# Stop the background daemon
roborev daemon stop

# Show daemon runtime info (PID, address, port)
roborev daemon info
```
```

--------------------------------

### Route Claude Code to a Proxy

Source: https://github.com/roborev-dev/roborev/blob/main/README.md

Configure the 'claude-code' agent to use a local proxy like Ollama for reviews by specifying a model and base URL. This example uses a local Ollama instance for reviews and Anthropic's Sonnet for fixes.

```toml
# .roborev.toml — local Ollama for reviews, real Anthropic for fixes
agent = "claude-code"
review_model = "glm-5.1:cloud@http://127.0.0.1:11434"
fix_model    = "sonnet"
```

--------------------------------

### Perform a Full-Scope Re-review (Since a Commit)

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-refine/SKILL.md

Executes an explicit full-scope re-review starting from a specified commit. Use --wait to block until the review is complete.

```bash
roborev review --since <commit> --wait
```

--------------------------------

### Build and Test Commands

Source: https://github.com/roborev-dev/roborev/blob/main/CLAUDE.md

Standard Go commands for building and testing the project. Use specific tags for different test types like integration or postgres.

```bash
go build ./...                       # Build
```

```bash
go test ./...                        # Test (unit tests only)
```

```bash
go test -tags integration ./...      # Test (unit + integration)
```

```bash
go test -tags postgres ./...         # Test (requires TEST_POSTGRES_URL)
```

--------------------------------

### Create Temporary Git Repositories for Tests

Source: https://context7.com/roborev-dev/roborev/llms.txt

Use these functions to set up temporary Git repositories for testing purposes. `NewTestRepo` creates a bare temporary directory with `git init`, while `NewTestRepoWithCommit` adds an initial commit.

```go
dir := testutil.NewTestRepo(t)
// bare temp dir + git init
```

```go
dir := testutil.NewTestRepoWithCommit(t)
// + initial commit
```

--------------------------------

### Run Project Tests

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-fix/SKILL.md

After applying fixes, execute the project's test suite to verify the changes. Replace 'go test ./...' with the appropriate command for your project if it differs.

```bash
go test ./...
```

--------------------------------

### Get Sync Status

Source: https://context7.com/roborev-dev/roborev/llms.txt

Retrieves the current status of the synchronization process.

```APIDOC
## GET /api/sync/status

### Description
Retrieves the current status of the synchronization process.

### Method
GET

### Endpoint
/api/sync/status
```

--------------------------------

### Get Health Info

Source: https://context7.com/roborev-dev/roborev/llms.txt

Retrieves health information about the Roborev service.

```APIDOC
## GET /api/health

### Description
Retrieves health information about the Roborev service.

### Method
GET

### Endpoint
/api/health
```

--------------------------------

### Get Daemon Status

Source: https://context7.com/roborev-dev/roborev/llms.txt

Retrieves the current status of the Roborev daemon.

```APIDOC
## GET /api/status

### Description
Retrieves the current status of the Roborev daemon.

### Method
GET

### Endpoint
/api/status
```

--------------------------------

### Get Review

Source: https://github.com/roborev-dev/roborev/blob/main/CLAUDE.md

Retrieves review details for a given job ID.

```APIDOC
## GET /review

### Description
Get review by job_id

### Method
GET

### Endpoint
/review
```

--------------------------------

### Set Up Full Worker Pool Test Context

Source: https://context7.com/roborev-dev/roborev/llms.txt

Creates a comprehensive test context including a worker pool, database, and agents. This allows for testing job creation, claiming, and retry exhaustion scenarios.

```go
ctx := newWorkerTestContext(t, 4 /* workers */)
job := ctx.createAndClaimJob()
job := ctx.createAndClaimJobWithAgent("codex")
ctx.exhaustRetries(job.ID) // drive job to failover state
```

--------------------------------

### Get Sync Status

Source: https://github.com/roborev-dev/roborev/blob/main/CLAUDE.md

Retrieves the current status of the PostgreSQL synchronization process.

```APIDOC
## GET /sync/status

### Description
Sync status

### Method
GET

### Endpoint
/sync/status
```

--------------------------------

### Get Job Output

Source: https://github.com/roborev-dev/roborev/blob/main/CLAUDE.md

Retrieves the accumulated output lines for a specific job.

```APIDOC
## GET /job/output

### Description
Accumulated output lines

### Method
GET

### Endpoint
/job/output
```

--------------------------------

### Get Fix Job Patch

Source: https://github.com/roborev-dev/roborev/blob/main/CLAUDE.md

Retrieves the patch generated by a completed fix job.

```APIDOC
## POST /job/patch

### Description
Get patch for completed fix job

### Method
POST

### Endpoint
/job/patch
```

--------------------------------

### Get sync status

Source: https://context7.com/roborev-dev/roborev/llms.txt

Check the current status of the synchronization process. The output is formatted using `jq`.

```shell
curl http://127.0.0.1:7373/api/sync/status | jq .
```

--------------------------------

### Initial Review Command

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-refine/SKILL.md

Execute an initial review using either `--since <commit>` to review commits after a specific commit, or `--branch` to review the current branch relative to its merge-base. The `--wait` flag is crucial as it exits with code 1 on failure, which is expected and should be handled.

```bash
roborev review --since <commit> --wait
```

```bash
roborev review --branch --wait
```

--------------------------------

### Get health info

Source: https://context7.com/roborev-dev/roborev/llms.txt

Fetch health information for the Roborev daemon. The response is piped to `jq` for pretty-printing.

```shell
curl http://127.0.0.1:7373/api/health | jq .
```

--------------------------------

### Execute Agent Review in Go

Source: https://context7.com/roborev-dev/roborev/llms.txt

Runs a review using a configured agent. The output is streamed to the provided io.Writer. Requires context, repository path, commit SHA, prompt, and a writer.

```go
// Run a review (streams output to w):
output, err := a.Review(ctx, "/path/to/repo", "HEAD", promptString, w)
```

--------------------------------

### Get Daemon Status

Source: https://github.com/roborev-dev/roborev/blob/main/CLAUDE.md

Retrieves the current status of the daemon, including version, queue statistics, and worker information.

```APIDOC
## GET /status

### Description
Daemon status (version, queue stats, workers)

### Method
GET

### Endpoint
/status
```

--------------------------------

### Queue a review with roborev

Source: https://context7.com/roborev-dev/roborev/llms.txt

Queue various git references for AI review. Reviews run asynchronously by default, but can be run locally or waited upon.

```bash
# Review HEAD (most recent commit)
roborev review
```

```bash
# Review a specific commit by SHA
roborev review abc1234
```

```bash
# Review a commit range (inclusive, start^..end)
roborev review abc1234 def5678
```

```bash
# Review all commits on the current branch since it diverged from main
roborev review --branch
```

```bash
# Review all commits on a specific branch
roborev review --branch=feature/my-feature
```

```bash
# Review all commits since a specific point (exclusive)
roborev review --since HEAD~5
```

```bash
# Review uncommitted (dirty) changes
roborev review --dirty
```

```bash
# Review with a security-focused system prompt
roborev review --type security
```

```bash
# Review with a design-focused system prompt
roborev review --type design HEAD
```

```bash
# Review and wait for the result in-terminal
roborev review --wait
```

```bash
# Use a specific agent and model
roborev review --agent claude-code --model claude-opus-4-5
```

```bash
# Use fast reasoning (lower latency)
roborev review --fast
```

```bash
# Run locally without the daemon (streams directly to stdout)
roborev review --local
```

```bash
# Enqueue dirty review via daemon API (HTTP equivalent)
curl -s -X POST http://127.0.0.1:7373/api/enqueue \
  -H 'Content-Type: application/json' \
  -d '{"repo_path":"/home/user/my-repo","git_ref":"dirty","diff_content":"<patch>","agent":"claude-code"}'
# => 201 Created: {"id":42,"agent":"claude-code","status":"queued",...}
```

--------------------------------

### Run Go PostgreSQL Tests

Source: https://context7.com/roborev-dev/roborev/llms.txt

Command to run tests specifically targeting PostgreSQL functionality. Requires the `TEST_POSTGRES_URL` environment variable to be set.

```bash
go test -tags=postgres -v ./internal/storage/...
```

--------------------------------

### Get daemon status

Source: https://context7.com/roborev-dev/roborev/llms.txt

Retrieve the current status of the Roborev daemon. The output is piped to `jq` for formatted JSON display.

```shell
curl http://127.0.0.1:7373/api/status | jq .
```

--------------------------------

### Invoke Roborev Design Review

Source: https://github.com/roborev-dev/roborev/blob/main/skills/roborev-design-review.md

Use this command to initiate a design review. Provide a file path to a specific design document or a job ID to fetch design output from roborev. If no argument is provided, the skill defaults to scanning the 'docs/design/' directory.

```bash
/roborev-design-review <path-or-job-id>
```

--------------------------------

### Test Helper Functions

Source: https://github.com/roborev-dev/roborev/blob/main/CLAUDE.md

Key test helper functions for creating temporary git repositories, managing commits, and interacting with the storage and daemon components.

```go
testutil.NewTestRepo()
```

```go
testutil.NewTestRepoWithCommit()
```

```go
testutil.InitTestRepo(t)
```

```go
testutil.GetHeadSHA(t, dir)
```

```go
openTestDB(t)
```

```go
createJobChain(t, db, path, sha)
```

```go
createRepo(t, db, path)
```

```go
createCommit(t, db, repoID, sha)
```

```go
claimJob(t, db, workerID)
```

```go
newWorkerTestContext(t, n)
```

```go
workerTestContext.createAndClaimJob()
```

```go
workerTestContext.createAndClaimJobWithAgent()
```

```go
workerTestContext.exhaustRetries()
```

```go
NewStaticConfig(cfg)
```

--------------------------------

### List Repositories

Source: https://github.com/roborev-dev/roborev/blob/main/CLAUDE.md

Lists all registered repositories.

```APIDOC
## GET /repos

### Description
List registered repos

### Method
GET

### Endpoint
/repos
```

--------------------------------

### Discover Open Reviews

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-fix/SKILL.md

If no job IDs are provided and no findings are present in the conversation, use this command to discover and list all open reviews. This command prints one line per open job, including its ID, commit SHA, agent, and summary.

```bash
roborev fix --open --list
```

--------------------------------

### Construct Roborev Review Command

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-design-review-branch/SKILL.md

Build the command to initiate a design review. The `--wait` flag ensures the command waits for completion before returning, and `--type design` specifies the review type. The `--base` flag is optional and used to specify a base branch for comparison.

```bash
roborev review --branch --wait --type design [--base <branch>]
```

--------------------------------

### Enqueue a fix job

Source: https://context7.com/roborev-dev/roborev/llms.txt

Use this endpoint to start a background worktree fix job, which applies a patch to the database. Requires agentic mode to be enabled.

```shell
curl -X POST http://127.0.0.1:7373/api/job/fix \
  -H 'Content-Type: application/json' \
  -d '{"parent_job_id":42,"agentic":true}'
```

--------------------------------

### Register a repository by path

Source: https://context7.com/roborev-dev/roborev/llms.txt

Register a new repository with Roborev by providing its local file system path. This operation requires POST and JSON content type.

```shell
curl -X POST http://127.0.0.1:7373/api/repos/register \
  -H 'Content-Type: application/json' \
  -d '{"path":"/home/user/my-repo"}'
```

--------------------------------

### Trigger immediate PostgreSQL sync

Source: https://context7.com/roborev-dev/roborev/llms.txt

Initiate an immediate synchronization with the PostgreSQL database. This is a POST request to the `/api/sync/now` endpoint.

```shell
curl -X POST http://127.0.0.1:7373/api/sync/now
```

--------------------------------

### Fetch batch of job+review pairs

Source: https://context7.com/roborev-dev/roborev/llms.txt

Retrieve a batch of job and review data by providing a list of job IDs. The API returns up to 100 IDs. The output is piped to `jq` for pretty-printing.

```shell
curl -X POST http://127.0.0.1:7373/api/jobs/batch \
  -H 'Content-Type: application/json' \
  -d '{"job_ids":[42,43,44]}' | jq .
```

--------------------------------

### Register Repository by Path

Source: https://context7.com/roborev-dev/roborev/llms.txt

Registers a new repository with Roborev by providing its file system path.

```APIDOC
## POST /api/repos/register

### Description
Registers a new repository with Roborev by providing its file system path.

### Method
POST

### Endpoint
/api/repos/register

### Request Body
- **path** (string) - Required - The file system path to the repository.
```

--------------------------------

### HTTP API - Get Review Output

Source: https://context7.com/roborev-dev/roborev/llms.txt

Retrieves the review output for a specific job ID. This endpoint is used to fetch the detailed results of a code review.

```APIDOC
## GET /api/review

### Description
Gets the review output for a specific job.

### Method
GET

### Endpoint
`/api/review`

### Query Parameters
- **job_id** (integer) - Required - The ID of the job to retrieve output for.

### Response
#### Success Response (200)
- **output** (string) - The review output content.
```

--------------------------------

### Construct Roborev Review Command

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-review-branch/SKILL.md

Build the command to initiate a Roborev review for a branch. Options for base branch and review type can be included.

```bash
roborev review --branch --wait [--base <branch>] [--type <type>]
```

--------------------------------

### Construct Roborev Design Review Command

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-design-review/SKILL.md

Build the command to initiate a design review for a specific commit. The `--wait` flag ensures the command waits for completion, and `--type design` specifies the review type. The commit can be omitted to default to HEAD.

```bash
roborev review [commit] --wait --type design
```

--------------------------------

### Open In-Memory SQLite Database for Storage Tests

Source: https://context7.com/roborev-dev/roborev/llms.txt

Provides an in-memory SQLite database instance suitable for testing storage-related functionalities.

```go
db := openTestDB(t)
```

--------------------------------

### HTTP API: Get a specific job by ID

Source: https://context7.com/roborev-dev/roborev/llms.txt

Retrieve details for a specific job by its ID using the `/api/jobs` endpoint with an `id` query parameter. The response contains job details in the `jobs` array.

```bash
curl "http://127.0.0.1:7373/api/jobs?id=42" | jq .jobs[0]

```

--------------------------------

### Trigger Immediate Sync via API

Source: https://context7.com/roborev-dev/roborev/llms.txt

Triggers an immediate synchronization process by sending a POST request to the /api/sync/now endpoint. Useful for ensuring data consistency on demand.

```bash
# Trigger immediate sync
curl -X POST http://127.0.0.1:7373/api/sync/now
```

--------------------------------

### Verify and Consolidate Review Findings

Source: https://context7.com/roborev-dev/roborev/llms.txt

Use `roborev compact` to have an agent verify, discard false positives, and consolidate open review findings into a single new job. Source jobs are closed upon success. Options include waiting for results, dry runs, specifying branches, limiting the number of jobs, and selecting agents.

```bash
roborev compact
```

```bash
roborev compact --wait
```

```bash
roborev compact --dry-run
```

```bash
roborev compact --branch main
```

```bash
roborev compact --all-branches
```

```bash
roborev compact --limit 15
```

```bash
roborev compact --agent claude-code --reasoning thorough
```

--------------------------------

### List Registered Repositories

Source: https://context7.com/roborev-dev/roborev/llms.txt

Lists all repositories that have been registered with Roborev.

```APIDOC
## GET /api/repos

### Description
Lists all repositories that have been registered with Roborev.

### Method
GET

### Endpoint
/api/repos
```

--------------------------------

### Respond to Review

Source: https://github.com/roborev-dev/roborev/blob/main/skills/README.md

Add a response to a review, documenting the changes made. This command is used after the issues have been fixed.

```bash
/roborev-respond 1019 Fixed null check and improved error handling
```

--------------------------------

### Register Repository

Source: https://github.com/roborev-dev/roborev/blob/main/CLAUDE.md

Registers a new repository by its path.

```APIDOC
## POST /repos/register

### Description
Register repo by path

### Method
POST

### Endpoint
/repos/register
```

--------------------------------

### Roborev CLI Commands

Source: https://github.com/roborev-dev/roborev/blob/main/CLAUDE.md

Core Roborev CLI commands for initialization, status checks, daemon operations, and the terminal UI.

```bash
make install                         # Install to ~/.local/bin
```

```bash
make lint                            # golangci-lint with --fix
```

```bash
roborev init                         # Initialize in a repo
```

```bash
roborev status                       # Check daemon/queue
```

```bash
roborev daemon run                   # Start daemon in foreground
```

```bash
roborev tui                          # Terminal UI
```

--------------------------------

### roborev-refine Skill Usage

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/codex/roborev-refine/SKILL.md

Command-line arguments for the roborev-refine skill. Use `--since` to specify a commit range and `--max-iterations` to set the loop limit.

```bash
$roborev-refine [--since <commit>] [--branch <name>] [--max-iterations <n>]
```

--------------------------------

### Run Roborev Review in Background

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-review/SKILL.md

Launches the roborev review command as a background task. This allows the user to continue working while the review is in progress. Use the Task tool with `run_in_background: true` and `subagent_type: "Bash"`.

```bash
roborev review [commit] --wait [--type <type>]
```

--------------------------------

### Subscribe to real-time SSE events

Source: https://context7.com/roborev-dev/roborev/llms.txt

Connect to the Server-Sent Events (SSE) stream to receive real-time notifications about events, such as job completions. Use `-N` to disable buffering.

```shell
curl -N http://127.0.0.1:7373/api/stream/events
```

--------------------------------

### Fetch Review Findings by Job ID

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-fix/SKILL.md

When job IDs are provided and findings are not already in the conversation, use this command to fetch the full review details as JSON for a specific job ID.

```bash
roborev show --job <job_id> --json
```

--------------------------------

### HTTP API: Queue a review job

Source: https://context7.com/roborev-dev/roborev/llms.txt

The `/api/enqueue` endpoint queues a review job. It requires a JSON payload specifying repository path, git ref, branch, agent, model, reasoning, and review type.

```bash
curl -X POST http://127.0.0.1:7373/api/enqueue \
  -H 'Content-Type: application/json' \
  -d '{
    "repo_path": "/home/user/my-repo",
    "git_ref": "HEAD",
    "branch": "feature/auth",
    "agent": "claude-code",
    "model": "claude-sonnet-4-5",
    "reasoning": "thorough",
    "review_type": ""
  }'
# 201: {"id":42,"agent":"claude-code","status":"queued",...}

```

--------------------------------

### Record Comments and Close Reviews

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-fix/SKILL.md

Use these separate commands to record a summary comment for a fixed job and then close the review. Ensure the comment command succeeds before running the close command. Escape quotes and special characters in the bash command.

```bash
roborev comment --commenter roborev-fix --job <job_id> "<summary of changes>"
# Only if the comment above succeeded:
roborev close <job_id>
```

--------------------------------

### Run Structured Code Analysis

Source: https://context7.com/roborev-dev/roborev/llms.txt

The `roborev analyze` command runs predefined analysis templates on specified files or branch changes. Use `--list` to see available types, `--fix` to auto-fix issues, and `--wait` to block until results are ready. Analysis jobs can be configured to run per file or output as JSON.

```bash
roborev analyze --list
```

```bash
roborev analyze duplication cmd/roborev/*.go
```

```bash
roborev analyze architecture internal/storage/
```

```bash
roborev analyze dead-code ./...
```

```bash
roborev analyze complexity --wait main.go
```

```bash
roborev analyze refactor --branch
```

```bash
roborev analyze refactor --fix ./...
```

```bash
roborev analyze complexity --per-file --wait internal/storage/*.go
```

```bash
roborev analyze duplication --agent gemini --json internal/*.go
```

```bash
roborev analyze --show-prompt test-fixtures
```

--------------------------------

### Resolve and Configure Agent Workflow in Go

Source: https://context7.com/roborev-dev/roborev/llms.txt

Resolves an agent configuration based on repository and global settings for a specific workflow. Supports CLI overrides for agent selection and reasoning level.

```go
// Reasoning levels (from fastest to most thorough):
// ReasoningFast, ReasoningStandard, ReasoningMedium, ReasoningThorough, ReasoningMaximum

// Resolve + configure an agent for a workflow (respects repo + global config):
resolution, err := agent.ResolveWorkflowConfig(
    "claude-code",   // CLI override (empty = use config)
    "/path/to/repo",
    cfg,             // *config.Config from config.LoadGlobal()
    "review",        // workflow: "review", "fix", "refine", "security", "design"
    "thorough",      // reasoning level string
)
// resolution.PreferredAgent, resolution.BackupAgent, resolution.ModelForSelectedAgent(...)
```

--------------------------------

### Create Isolated Worktree in Go

Source: https://context7.com/roborev-dev/roborev/llms.txt

Creates a detached-HEAD worktree in an isolated directory for running agents without modifying the user's working tree. Ensure to defer closing the worktree to clean up temporary files.

```go
// Create an isolated detached-HEAD worktree at HEAD:
wt, err := worktree.Create("/path/to/repo", "HEAD")
defer wt.Close() // removes the temporary worktree
```

--------------------------------

### roborev-respond Command Usage

Source: https://github.com/roborev-dev/roborev/blob/main/internal/skills/claude/roborev-respond/SKILL.md

This is the basic command structure for the roborev-respond skill. It requires a job ID and can optionally take a message.

```bash
/roborev-respond <job_id> [message]
```