actual

name	actual
description	Feature-complete companion for the actual CLI, an ADR-powered CLAUDE.md/AGENTS.md generator. Runs and troubleshoots actual adr-bot, status, auth, config, runners, and models, and drives the Actual Advisor (login, advisor, whoami, logout) for org-scoped architecture questions. Covers all 5 runners (claude-cli, anthropic-api, openai-api, codex-cli, cursor-cli), all model patterns, all 3 output formats (claude-md, agents-md, cursor-rules), and all error types. Use when working with the actual CLI, running actual adr-bot, configuring runners or models, troubleshooting errors, managing output files, or asking the Advisor what your team's ADRs say about a design choice.
argument-hint	[question or command] e.g. 'run sync', 'set up anthropic-api', 'fix ClaudeNotFound'
metadata	{"openclaw":{"requires":{"bins":["actual"]},"primaryEnv":"ANTHROPIC_API_KEY","homepage":"https://cli.actual.ai"}}

actual CLI Companion

Inline knowledge and operational workflows for the actual CLI. Read this file first; load reference files only when you need deeper detail for a specific topic.

CLI Not Installed

If the actual binary is not in PATH, stop and help the user install it before doing anything else. All commands, pre-flight checks, and diagnostics require the CLI.

Detect with:

command -v actual

Install options (try in this order):

Method	Command
npm/npx (quickest)	npm install -g @actualai/actual
Homebrew (macOS/Linux)	brew install actual-software/actual/actual
GitHub Release (manual)	Download from actual-software/actual-releases on GitHub

For one-off use without installing globally:

npx @actualai/actual adr-bot [flags]

After install, verify: actual --version

ADR Pre-Check (OpenClaw)

Before creating a new skill, component, or feature, check whether the project has ADR context available. This ensures new work aligns with existing architectural decisions.

1. Check for managed section markers in CLAUDE.md or AGENTS.md:

   grep -l "managed:actual-start" CLAUDE.md AGENTS.md 2>/dev/null
   

2. If markers found: ADR context is already loaded into the output file. Proceed — the agent session already has this context.
3. If no markers but actual CLI is available:

   actual adr-bot --dry-run
   

   Review the output. If relevant ADRs exist, run actual adr-bot to sync them into CLAUDE.md / AGENTS.md before starting work.
4. If no markers and no CLI: proceed without. Do not block the user.

Commands

Command	Purpose	Key Flags
actual adr-bot	Analyze repo, fetch ADRs, tailor, write output	--dry-run [--full], --force, --no-tailor, --project PATH, --model, --runner, --verbose, --reset-rejections, --max-budget-usd, --no-tui, --output-format, --show-errors
actual status	Check output file state	--verbose
actual auth	Check authentication status	(none)
actual config show	Display current config	(none)
actual config set <key> <value>	Set a config value	(none)
actual config path	Print config file path	(none)
actual runners	List available runners	(none)
actual models	List known models by runner	(none)
actual login	Sign in to Actual AI (required for the Advisor)	--no-browser, --org, --api-url
actual advisor "<question>"	Ask an org-scoped architecture question	--repo, --org, --api-url
actual whoami	Show the signed-in identity (no network call)	(none)
actual logout	Sign out and clear local credentials	(none)

login / advisor / whoami / logout drive the Actual Advisor — a signed-in, org-scoped Q&A surface separate from the runner auth that auth reports. See Asking the Advisor.

Runner Decision Tree

Use this to determine which runner a user needs:

Has claude binary installed?
  YES -> claude-cli (default runner, no API key needed)
  NO  -> Do they want Anthropic models?
           YES -> anthropic-api (needs ANTHROPIC_API_KEY)
           NO  -> Do they want OpenAI models?
                    YES -> codex-cli or openai-api (needs OPENAI_API_KEY)
                    NO  -> cursor-cli (needs agent binary, optional CURSOR_API_KEY)

Runner Summary

Runner	Binary	Auth	Default Model
claude-cli	claude	claude auth login	claude-sonnet-4-6
anthropic-api	(none)	ANTHROPIC_API_KEY	claude-sonnet-4-6
openai-api	(none)	OPENAI_API_KEY	gpt-5.2
codex-cli	codex	OPENAI_API_KEY or codex login (ChatGPT OAuth)	gpt-5.2
cursor-cli	agent	Optional CURSOR_API_KEY	(cursor default)

Model-to-Runner Inference

The CLI auto-selects a runner from the model name:

Model Pattern	Inferred Runner
sonnet, opus, haiku (short aliases)	claude-cli
claude-* (full names)	anthropic-api
gpt-*, o1*, o3*, o4*, chatgpt-*	codex-cli
codex-*, gpt-*-codex*	codex-cli
Unrecognized	Error with suggestions

For deep runner details (install steps, compatibility, special behaviors), see references/runner-guide.md.

Sync Quick Reference

The most common sync patterns:

# Preview what sync would do (safe, no file changes)
actual adr-bot --dry-run

# Preview with full content
actual adr-bot --dry-run --full

# Run sync, skip confirmation prompts
actual adr-bot --force

# Sync specific subdirectories only (monorepo)
actual adr-bot --project services/api --project services/web

# Use a specific runner/model
actual adr-bot --runner anthropic-api --model claude-sonnet-4-6

# Skip AI tailoring (use raw ADRs)
actual adr-bot --no-tailor

# Re-offer previously rejected ADRs
actual adr-bot --reset-rejections

# Set spending cap
actual adr-bot --max-budget-usd 5.00

For the complete 13-step sync internals, see references/sync-workflow.md.

Operational Workflow: Running Sync

Follow this pattern whenever running sync. Do NOT skip pre-flight.

0. Verify CLI installed (LOW freedom -- exact check)

command -v actual       # Must succeed before anything else

If missing, follow the install steps in CLI Not Installed above. Do NOT proceed until actual --version succeeds.

1. Pre-flight (LOW freedom -- exact commands)

actual runners          # Verify runner is available
actual auth             # Verify authentication (for claude-cli)
actual config show      # Review current configuration

If any check shows a problem, diagnose and fix before proceeding.

2. Dry-run (LOW freedom -- exact command)

actual adr-bot --dry-run [--full] [user's flags]

Show the user what would change. Let them review.

3. Confirm (HIGH freedom)

Ask user if they want to proceed. If no, stop.

4. Execute (LOW freedom -- exact command)

actual adr-bot [user's flags]

5. On failure: Diagnose

Match the error against the troubleshooting table below. For full error details, load references/error-catalog.md.

6. Fix and retry

Apply the fix, then return to step 1 to verify.

Operational Workflow: Diagnostics

For comprehensive environment checks, run the bundled diagnostic script:

bash .claude/skills/actual/scripts/diagnose.sh

This checks all binaries, auth status, environment variables, config, and output files in one pass. It is read-only and never modifies anything.

Use inline commands instead when checking a single thing (e.g., just actual auth).

Asking the Advisor

Beyond generating guardrails, the actual CLI can answer architecture questions on demand. The advisor command queries your organization's ADRs and returns guidance grounded in your team's real decisions — reach for it when the user asks "what's our convention for X?" or "does our architecture allow Y?" instead of answering from general knowledge.

The Advisor needs a signed-in Actual AI account, which is separate from the runner auth that actual auth reports. The flow is sign in, ask, then sign out when done:

# 1. Sign in via browser OAuth (grants the adr:query and adr:review scopes).
actual login
#    Headless or over SSH? Print the URL instead of opening a browser:
actual login --no-browser

# 2. Ask an org-scoped architecture question.
actual advisor "Which router should new Next.js pages use?"

# 3. Check identity, or sign out when finished.
actual whoami
actual logout

The Advisor prints the answer followed by the related ADRs it drew on, each with a confidence score. Scope a question to one connected repository with --repo <UUID> or to a specific organization with --org.

Both login and advisor default to the production Actual AI endpoints out of the box (https://app.actual.ai for sign-in, the prod api-service for queries), so no URL flag is needed for normal use. Pass --api-url, or set ACTUAL_AUTH_URL / ACTUAL_API_URL, only to target staging or a local mock. whoami reads the local credentials and makes no network call (it errors when signed out); logout revokes the session server-side on a best-effort basis and then always clears the local credentials, even offline.

Troubleshooting Quick Reference

Error	Exit Code	Likely Cause	Quick Fix
ClaudeNotFound	2	claude binary not in PATH	Install Claude Code CLI
ClaudeNotAuthenticated	2	Not logged in	Run claude auth login
CodexNotFound	2	codex binary not in PATH	Install Codex CLI
CodexNotAuthenticated	2	No auth for codex	Set OPENAI_API_KEY or run codex login
CursorNotFound	2	agent binary not in PATH	Install Cursor CLI
ApiKeyMissing	2	Required env var not set	Set ANTHROPIC_API_KEY or OPENAI_API_KEY
CodexCliModelRequiresApiKey	2	ChatGPT OAuth with explicit model	Set OPENAI_API_KEY (OAuth only supports default model)
CreditBalanceTooLow	3	Insufficient API credits	Add credits to account
ApiError	3	API request failed	Check API URL, network, credentials
ApiResponseError	3	Unexpected API response	Check API status, retry
RunnerFailed	1	Runner process errored	Check runner output, logs
RunnerOutputParse	1	Could not parse runner output	Check model compatibility
RunnerTimeout	1	Runner exceeded time limit	Increase invocation_timeout_secs
ConfigError	1	Invalid config file	Check YAML syntax, run actual config show
AnalysisEmpty	1	No analysis results	Check project path, repo content
TailoringValidationError	1	Tailored output invalid	Retry, or use --no-tailor
IoError	5	File I/O failure	Check permissions, disk space
UserCancelled	4	User cancelled operation	(intentional)

For full error details with hints and diagnosis steps, see references/error-catalog.md.

Exit Code Categories

Code	Category	Errors
1	General / runtime	RunnerFailed, RunnerOutputParse, ConfigError, RunnerTimeout, AnalysisEmpty, TailoringValidationError, InternalError, TerminalIOError
2	Auth / setup	ClaudeNotFound, ClaudeNotAuthenticated, CodexNotFound, CodexNotAuthenticated, CursorNotFound, ApiKeyMissing, CodexCliModelRequiresApiKey
3	Billing / API	CreditBalanceTooLow, ApiError, ApiResponseError
4	User cancelled	UserCancelled
5	I/O	IoError

Config Quick Reference

Config file: ~/.actualai/actual/config.yaml (override with ACTUAL_CONFIG or ACTUAL_CONFIG_DIR env vars).

Most-used config keys:

Key	Default	Purpose
runner	claude-cli	Which runner to use
model	claude-sonnet-4-6	Model for Anthropic runners
output_format	claude-md	Output format: claude-md, agents-md, cursor-rules
batch_size	15	ADRs per batch (min 1)
concurrency	10	Parallel requests (min 1)
invocation_timeout_secs	600	Runner timeout in seconds (min 1)
max_budget_usd	(none)	Spending cap (positive, finite)

For all 18 config keys with validation rules, see references/config-reference.md.

Output Formats

Format	File	Header
claude-md (default)	CLAUDE.md	# Project Guidelines
agents-md	AGENTS.md	# Project Guidelines
cursor-rules	.cursor/rules/actual-policies.mdc	YAML frontmatter (alwaysApply: true)

Managed sections use markers: <!-- managed:actual-start --> / <!-- managed:actual-end -->.

Merge behavior:

• New root file: header + managed section
• New subdir file: managed section only (no header)
• Existing with markers: replace between markers, preserve surrounding content
• Existing without markers: append managed section

For full format details and merge internals, see references/output-formats.md.

Reference Files

Load these only when you need deeper detail on a specific topic:

File	When to Load
references/sync-workflow.md	Debugging sync failures, understanding sync internals
references/runner-guide.md	Setting up a runner, model compatibility, runner-specific behavior
references/error-catalog.md	Troubleshooting a specific error with full diagnosis steps
references/config-reference.md	Looking up config keys, validation rules, dotpath syntax
references/output-formats.md	Output format questions, managed section behavior, merge logic

Additional Resources

For anything not covered by this skill or its reference files, fetch the full CLI documentation:

• Full docs (Markdown): https://cli.actual.ai/docs.md — complete command reference, all flags, runners, output formats, config keys, and troubleshooting
• LLM summary: https://cli.actual.ai/llms.txt — concise machine-readable overview