dax

README

dax

A coding agent for the terminal. It reads, searches, edits and writes files in your project and runs commands, asking you before it does anything it should not do alone, and records every session so you can list, resume, verify and export it. It runs against a local Ollama with no setup, OpenAI, Anthropic or Gemini with a key, or Claude and Gemini on Google Vertex AI with your gcloud login.

It is a thin product over a set of small Go libraries (openresponses, agentturn, agentkit, agentpolicy, agentsession and friends); docs/design.md says who owns what.

dax is a minimal core meant to be extended. The session (the model, the store, the policy, the workspace, AGENTS.md, MCP servers, compaction) knows no tool. Everything the model can do comes from an extension, and dax's own are extensions like any other: dax-coding (the file tools and bash), dax-agents (the sub-agents), dax-skills and dax-memory. A Go program runs dax with extensions of its own on the same terms, or leaves one of dax's out; see Building on dax.

Where the tools act is a value too. dax is meant to feel the same whether its tools act on this machine or elsewhere: every tool, and every check the policy makes of a call, goes through one workspace interface, and this machine's directory is one implementation of it; see Where the tools act.

Install

go install github.com/ChristopherDavenport/dax/cmd/dax@latest
# or, in a checkout
make install

Go 1.26 or later. With no flags dax talks to Ollama on localhost:11434 and runs qwen3.5:9b.

dax                          # REPL in the current directory; Ctrl-C aborts a run in flight
dax -p "what is in go.mod"   # one prompt, then exit
dax -provider anthropic      # a hosted model

Providers and keys

-provider	default model	key (environment or api_key_command)	-base-url
ollama (default)	qwen3.5:9b	none	default http://localhost:11434/v1; another Ollama host
openai	gpt-5	OPENAI_API_KEY	not supported
openrouter	deepseek/deepseek-v4-pro-0813; sub-agents deepseek/deepseek-v4.1-flash	OPENROUTER_API_KEY	not supported
openresponses	none; -model is required	the variable -api-key-env names, or none	required
anthropic	claude-sonnet-5-5	ANTHROPIC_API_KEY	not supported
gemini	gemini-2.5-pro	GEMINI_API_KEY	not supported
vertex	claude-sonnet-5-5	none; Google Application Default Credentials	not supported

A provider is a vendor: openai is OpenAI, and openrouter is OpenRouter, whose models are named vendor/model. Any other server that speaks the Open Responses (OpenAI Responses) API, such as vLLM, LM Studio or a proxy, is openresponses, with its -base-url, a -model, and -api-key-env naming the variable its key is in, if it takes one:

LITELLM_KEY=... dax -provider openresponses -base-url https://llm.internal/v1 -model qwen3-coder -api-key-env LITELLM_KEY

-model names another model. -subagent-model names the one the sub-agents run; without it a sub-agent runs the provider's default for sub-agents, where it has one, else the main model. A missing key is an error before any request is made: dax: openai: no API key: set OPENAI_API_KEY in the environment. dax reads keys from the environment and nowhere else (not from a config file, not from a flag, so they stay out of ps and out of a repository), and never prints one. The variable a key was read from is removed from the environment of bash commands and MCP servers, whatever it is called, unless pass_env names it.

For a key that expires, such as a short-lived access token, set api_key_command in your config (or -api-key-command) to a program that prints the key, in place of the variable:

{
  "provider": "openresponses",
  "base_url": "https://llm.internal/v1",
  "model": "qwen3-coder",
  "api_key_command": ["gcloud", "auth", "print-access-token"]
}

dax runs it without a shell when it starts, again once the key is five minutes old, and again when the server answers 401 or 403, then sends the refused request once more with the new key, so a long session outlives any one key. Its output must be a single token of printable ASCII, at most 16 KiB. It may take up to five minutes, long enough to wait on a sign-in in a browser. Concurrent requests share one run. The key is held in memory only: it is never written down or printed, and never put in any environment, so bash commands and MCP servers cannot see it. It is for the providers that take a key (not ollama or vertex), and is one source or the other with api_key_env. -api-key-command "" turns off a command your config sets.

When the command fails, or the server refuses even a fresh key, the turn ends with an error that starts authentication failed and carries the last line the command wrote to standard error. A failure is not retried and the command is not run again for a few seconds, so a helper that cannot sign in is not run over and over; send the prompt again once you have signed in. The REPL and -p show the command's standard error as it runs; the terminal client keeps it for the error, so it does not draw over the screen. To have the error say how to sign in, set api_key_login to a URL or a command:

  "api_key_command": ["my-token-helper", "print"],
  "api_key_login": "my-token-helper login"

gives authentication failed; sign in with my-token-helper login, then try again (...). dax only shows it; it does not open or run it.

A server may group a client's calls by session, or key its prompt cache on the session, and record which client called. session_header names the header each model call carries its session's ID in, and client_header the header that carries dax/ and dax's version:

  "session_header": "X-Session-Id",
  "client_header": "X-Client"

The ID is that of the session the call is recorded in: a sub-agent's calls carry the sub-agent's own session, and after /clear the new session's. A request made outside a run, such as asking the vendor what a model supports, carries none. Each must be a header name of letters, digits and hyphens, not one the client sets itself (Authorization, X-Api-Key, Content-Type and the like), and the two must differ. They are not for vertex.

Ollama, OpenAI, OpenRouter and openresponses use the openresponses client; Anthropic and Gemini use its provider adapters.

vertex is Google Vertex AI, which serves Claude and Gemini behind one project and login. Each request goes to the Anthropic adapter for a claude- model and to the Gemini adapter for a gemini- model, so the main agent and the sub-agents can run different families and /model can move between them; any other model name is refused. It takes no key: it authenticates with Application Default Credentials (gcloud auth application-default login, or GOOGLE_APPLICATION_CREDENTIALS). It reads Google Cloud's usual settings from the environment: the project from GOOGLE_CLOUD_PROJECT, falling back to the credentials' own project, and the location from GOOGLE_CLOUD_LOCATION (or GOOGLE_CLOUD_REGION), such as global or us-east5.

gcloud auth application-default login
export GOOGLE_CLOUD_PROJECT=my-project GOOGLE_CLOUD_LOCATION=global
dax -provider vertex -model claude-opus-5-5 -subagent-model gemini-3.5-flash

Vertex publishes no model capabilities, so what dax knows is read off the model ID's generation, from what Vertex accepted from each model:

models	efforts	so
Claude before 4.6 (claude-haiku-4-5, claude-opus-4-5)	none	reasoning is off; the adapter sends an effort as adaptive thinking, which they do not take
Claude 4.6	none to high	xhigh asks for high
Claude 4.7 to 5	not described	sent as configured
Claude 5.5 and later	minimal to xhigh	-think=false asks for minimal: they refuse the way the adapter turns thinking off
Gemini	low to high	none and minimal ask for low, xhigh for high

What the model takes

At start dax asks the vendor what the model supports and shows it under the banner, model: reasoning low–max, always on · default high · 1M context · 128k out (openrouter). Anthropic and Gemini are asked through their model endpoints, OpenRouter through its public catalogue (fetched once, without the key) and Ollama through /api/show; OpenAI and an openresponses server publish nothing, and nothing is shown.

The reasoning effort dax asks for is then fitted to the answer: an effort the model does not take becomes the nearest one it does, and dax says so once, [qwen3-coder:30b: reasoning effort low is not accepted; asking for none (ollama)]. So -think on a model that cannot reason turns reasoning off instead of failing, and -think=false on one that always reasons asks for its least effort. The fitting is done where each configuration is made (the main agent's, the sub-agents', a fold's summary, /think and /model), never to a request on its way out, so the session records the effort that was sent. When the vendor cannot be asked, or does not know the model, the effort is asked for as configured. -think (default on) asks for the -effort (minimal, low by default, medium, high or xhigh) with summaries and shows the reasoning; -think=false turns reasoning off. The effort is fitted like any other, so medium on a model that takes only low and high asks for low.

Sub-agents

The main agent can start sub-agents, each a run of its own on the sub-agent model, recorded as a child session:

• explore investigates without changing anything (read, glob, grep, ls and bash) and returns a written answer.

• task carries out a coding task with the file tools and bash, then reports what it changed, what it ran and what is left. Its instructions are dax's, your instructions_file and the AGENTS.md chain, without skills or memory. The main agent chooses, per call:

  • context: fresh (the default) sees only the brief, so the brief must be complete; fork also sees the conversation so far, for a task that depends on what was said or decided in it.
  • model: subagent (the default) runs the sub-agent model; main runs the main model, for a task that needs it.

  Each call reads /think and /model as they stand: main is the model /model last chose, and with no subagent_model configured the sub-agents run it too.

  A fork's conversation is recorded in the child session as its own opening items, so the child verifies like any session. On OpenRouter the defaults were the cheapest in a trial: a fresh Flash sub-agent cost about a twentieth of a Pro one at the same pass rate, and a Pro fork about twice a fresh Pro task.

Several calls in one turn run at once, so the main agent can hand out independent pieces of work in parallel. Writes and edits take a workspace lock, so two sub-agents editing one file cannot lose an edit, though they can still make changes that do not fit together; the sub-agent is told to change only the files its task is about.

Both are offered by default; -agents=false or "agents": false turns them off. They are the dax-agents extension, which ships an allow rule for starting one, since what a sub-agent then does is decided call by call by the same policy.

Config

Settings come from three layers, each overriding the one before:

1. ~/.config/dax/config.json ($XDG_CONFIG_HOME/dax/config.json; -config path names another)
2. .dax/config.json in the working directory, read through the workspace, which can only tighten the policy
3. flags

{
  "provider": "anthropic",
  "model": "claude-sonnet-5-5",
  "base_url": "",
  "think": true,
  "effort": "high",
  "instructions_file": "~/.config/dax/instructions.md",
  "skills_dirs": ["~/skills"],
  "memory_dir": "~/.dax/memory",
  "pricing_file": "~/.config/dax/prices.json",
  "mcp_servers": {
    "fs": { "command": "mcp-server-filesystem /home/me/notes" }
  },
  "policy": {
    "builtin": true,
    "fallback": "ask",
    "allow": ["bash(make:*)", "write(docs/**)"],
    "ask": ["bash(go test -race:*)"],
    "deny": ["bash(git push --force:*)"]
  }
}

model, subagent_model, base_url, api_key_env, api_key_command, api_key_login, session_header and client_header belong to the provider in force where they are set. A later layer that switches the provider leaves them behind, so with "model": "qwen3-coder:30b" in your config, dax -provider openrouter runs OpenRouter's default model rather than asking OpenRouter for an Ollama one; -model beside -provider names another.

Every field is optional. The files are strict: an unknown field, a bad provider, a malformed URL or a rule that does not parse is an error that names the file and the field.

field	meaning
provider, model, subagent_model, base_url, think, effort	as above; base_url is for ollama and openresponses
agents	offer the explore and task sub-agents; default true
api_key_env	the variable holding the openresponses provider's key (the name, never the key)
api_key_command	a program and its arguments, ["program", "arg", ...], that prints the provider's key; run again as the key ages or is refused (see above)
api_key_login	how to sign in again, a URL or a command, shown when api_key_command fails or its key is refused; one line, at most 500 bytes
session_header	the header each model call carries its session's ID in, for a server that groups calls by session (see above)
client_header	the header each model call carries dax/<version> in, for a server that records which client called
instructions_file	your own instructions, added to the system prompt after dax's; a relative path is relative to the file that names it
skills_dirs	more skill directories, after .dax/skills and ~/.dax/skills; one that does not exist is an error
agents_md_global	your own instruction files for every session, whatever the repository holds, such as an AGENTS.md in a directory above your checkouts: read on this machine after ~/.dax/AGENTS.md and before the repository's AGENTS.md chain, in order; a missing one is skipped; a relative path is relative to the file that names it
memory_dir	where the model's memory lives; "" turns memory off. Default ~/.dax/memory
pricing_file	a JSON file of model prices ("model": {"input","cached","output"} in USD per million tokens), used by the terminal client's status line and session pane to show cost; without it the client shows token usage but no cost
mcp_servers	stdio MCP servers by name; the name prefixes their tools, mcp__<name>__<tool>
max_read_bytes	the most bytes of a file read scans per call and edit will rewrite; default 2 MiB (use grep to find a line later in a bigger file)
pass_env	credential-looking variables bash commands and MCP servers may inherit, by name (default none)
policy	see below

A project's .dax/config.json comes from a repository, not from you, so it can only tighten. It may add ask and deny rules to policy (no allow, and no ! carve-out of any rule), set "builtin": false to drop the shipped allow rules (their asks and denies stay), and set "fallback" to ask or deny when that is stricter than yours. It cannot bring back what you dropped or loosen what you set, and its rules rank below yours so they cannot cancel one of yours. It may not set provider, model, subagent_model, base_url, api_key_env, api_key_command, api_key_login, session_header, client_header, think, effort, agents, instructions_file, skills_dirs, agents_md_global, memory_dir, pricing_file or mcp_servers: dax refuses the file with an error naming the field and saying to put it in your own config. (Where the model runs, what it is told and remembers, and what programs start are decisions that send your code, your files and your keys somewhere; a repository does not get to make them.)

flag
-provider, -model, -subagent-model, -base-url, -api-key-env, -think, -effort	override the config
-api-key-command 'program arg ...'	overrides api_key_command, split on spaces; "" turns it off, and api_key_login with it
-api-key-login hint	overrides api_key_login
-session-header name, -client-header name	override session_header and client_header; "" turns one off
-config path	the user config file
-memory dir	memory directory; off or empty disables it
-pricing-file path	JSON file of model prices for the terminal client's session cost
-no-policy	run every tool call without asking; ignores the config's policy
-front tui|repl	the front end; default tui when standard input and output are both terminals and a session is recorded, repl otherwise (so -sessions "" gives the REPL); -front tui without a terminal or a session store is refused; -p always prints
-v	the terminal client prints its start lines before it takes the screen and what dax noted during the run after it exits; without it, only warnings before and the resume command after
-agents-md-global 'a.md:b.md'	overrides agents_md_global, split on : as PATH is, so a path may hold a space; "" clears it
-agents-md, -skills, -trust-skills	the AGENTS.md files (~/.dax/AGENTS.md, agents_md_global and the repository's chain), skills, and a skill's allowed-tools running unasked until the next message, for skills in ~/.dax/skills and skills_dirs only, never the repository's
-compact N, -compact-server	fold the transcript above N estimated tokens, locally or through the server; without -compact, N is three quarters of the model's context window when the vendor reports the window, and compaction is off when it does not; -compact 0 turns it off
-mcp "cmd"	one more stdio MCP server, as mcp__cli__<tool>
-agents	offer the explore and task sub-agents (default on; -agents=false turns them off)
-sessions dir, -sync append|response|never	the session store (-sessions "" disables recording) and when appends are durable

Tools

tool	what it does
read	numbered lines of a file, with offset and limit
write	create or replace a file, making parent directories
edit	replace one exact occurrence of a string
glob	paths matching a doublestar pattern (**/*.go, cmd/*/main.go, *.{md,txt}), sorted
grep	a regular expression over files: path, include (a glob), ignore_case, max_results; prints path:line:text
ls	one directory, sorted, directories with /, files with their size
bash	a command in the working directory with a timeout; runs one at a time, its output reported as progress while it runs

The file tools are confined to the working directory. A path that is absolute outside it, climbs out with .., or goes out through a symbolic link is refused. bash is not confined (a shell reaches what you reach); the policy is what stands in front of it. glob and grep skip .git, node_modules, vendor, .venv, __pycache__ and similar trees by name; there is no .gitignore support. Search a skipped directory by naming it as path.

Environment of commands and servers

Commands the bash tool runs and the MCP servers dax starts do not get your credentials. A variable is removed from their environment if its name ends in _KEY, _KEY_ID, _PAT, _PWD, _JWT, _CREDENTIALS, _AUTH, _TOKEN, _SECRET, _PASSWORD or _API_KEY, contains PASSWORD or SECRET, is one of PASSWORD, TOKEN, API_KEY, SECRET_KEY, DATABASE_URL, SSH_AUTH_SOCK, PGPASSWORD, MYSQL_PWD or a provider or CI token by name (OPENAI_API_KEY, GITHUB_TOKEN, ...), or holds a URL with user:password@ in it. So a test, a build script or a server cannot read them, and cannot use your ssh agent.

Variables that hold the path of a credential file pass through, since a program that needs its file needs them: GOOGLE_APPLICATION_CREDENTIALS, KUBECONFIG, DOCKER_CONFIG, NETRC, AWS_SHARED_CREDENTIALS_FILE, AWS_CONFIG_FILE, CLOUDSDK_CONFIG, PGPASSFILE. The files are as readable to a command as they are to you. dax has no setting to withhold them; unset one before starting dax to keep it from commands.

A command that needs a scrubbed variable (gh, a private module proxy) gets it by name from your config: "pass_env": ["GITHUB_TOKEN"]. Only your own config can say that.

Policy

Each tool call is decided by rules: deny, then ask, then allow, then the fallback (ask). A rule is a tool name, or a name with a specifier: read, bash(go test:*), write(docs/**). Bash, Read and Edit are accepted as the reference's names. For bash the specifier is matched against the command; go test:* means go test and anything after it at a word boundary, so it matches go test ./... and not go testing.

dax ships this default, as the rules of its extensions (a verdict names the one whose rule decided it: extension:dax-coding, say):

• runs without asking: read, glob, grep, ls and the read-only bash commands listed below (dax-coding's), explore and task (dax-agents'), skill (dax-skills') and memory_search (dax-memory's);
• asks: write, edit, every other command, memory writes, and every MCP tool. go test, go build, go vet and go list ask: they run the repository's code (a TestMain, cgo, a vet tool), so a hostile repository would run code as you on the model's say-so.

What runs without asking is a safe subset, not a blacklist. A bash call is auto-allowed only when it parses in a strict subset of the syntax and every part of it is a command dax knows to be read-only, with arguments that are.

The syntax: words of letters, digits and _ . / : @ % + , = - (and ~ or ^ inside a word, as in HEAD~1), single-quoted strings, and double-quoted strings with no $, backtick or backslash; the operators && and |; at the end of a command 2>&1, 2>/dev/null or >/dev/null as a separate word; and * or ? in the arguments of ls. Anything else, ;, &, ||, <, any other >, #, $, backtick, backslash, parentheses, braces, [, a newline, = in the command word, a non-ASCII byte, asks. It is never denied for that.

The commands (&& joins any of them; after a | only head, tail, wc, sort, uniq, cut and grep, with flags from a short list and no file arguments):

• git status, diff, log, show, branch (listing forms only, a name only with --list), rev-parse, ls-files, remote (-v only), blame, stash list, tag (listing forms only), describe, shortlog and config --get of one key that holds no secret (user.name, user.email, core.autocrlf, branch.*.remote, remote.*.url, ...; not --list, --get-regexp, --global, --system, http.* or credential.*), with read-only flags from an allowlist, including combined short flags (-sb) and space-separated values (-n 5, --author x, -S foo). Not --output, -o, --ext-diff, --textconv, -c, -C, --git-dir, --no-index, or %G in a format.
• ls (listing flags; a glob is expanded and checked), pwd, go version, go env NAME.
• cat, head, tail, wc and grep (not recursive) on named files. Each file must be a regular file inside the working directory no larger than max_read_bytes (default 2 MiB), checked when the call is decided, and what bash returns is capped at 50 KiB however big it is; grep -r asks, use the grep tool.
• cd <directory inside the workspace> && ...: the directory is checked, and what follows is checked relative to it.

The arguments: every path stays inside the working directory with links resolved and no .. component at all, and no git revision or pathspec contains a : (HEAD:file and :/file name what is in the repository, which may be above the working directory).

git: before a git command runs unasked, dax asks git for the config it would use (git config --list --show-scope). If the repository's own config, or anything it includes, names a program (filter.*.clean, smudge or process, diff.*.textconv or command, core.askPass, editor, gitProxy, attributesFile, remote.*.uploadpack or receivepack, credential.helper, merge.*.driver, pager.*, protocol.*.allow = always, ...), moves the work tree (core.worktree), is bare, sets an extension, or runs a submodule update command, or if a submodule's own config names a program, the call asks and the question names the key. It also asks when .git is a file or a link, or when git's own answer for the repository (rev-parse) is not the workspace's ancestor chain: a hand-built .git can point git at another repository or at your home directory. Your own global and system config are trusted. The auto-allowed run switches off the rest: no fsmonitor, pager, ssh command or hooks, the gpg programs are /bin/false, --no-ext-diff --no-textconv are added to diff, log and show, and go runs with GOTOOLCHAIN=local. What an auto-allowed command prints has the user:token@ of any URL replaced by ***@. A command you approve runs as you would run it, hooks and GIT_CONFIG_* included.

What was decided is what runs: the policy stamps an auto-allowed call with the plan it approved, and the bash tool runs a stamped call only if the line still analyses to that plan. If a file or the repository's config changed in between, the call fails with "the command changed since it was allowed; ask again", and the original line is never run instead. Only dax can stamp a call; one the model stamps is refused. ls with a glob runs with the expansion after a --, so a file called -n is a name.

Secret-looking files ask (any case, and under any name: a link to one is decided under its target as well): reading .env*, *.pem, *.key, id_*, *.p12, *.pfx, .npmrc, .netrc, .pgpass, .git-credentials, credentials*, *secret*, .aws/**, .ssh/**, .kube/config or .docker/config.json (at the root or in any directory) asks, whether by read, grep, glob or ls or by cat, head, tail, wc, grep and git show, log, diff or blame of the path: the stage is decided as a read of it. Name the path to open one, in your own config: "allow": ["read(.env)"] (or Read(config/*.pem)) allows it for every tool, cat .env included, and only that. A bare "allow": ["read"] does not open them: an ask beats an allow. This is a guard on naming a file, not a boundary: a search of the whole tree (grep -r, the grep tool with no path) can still read one, and so can anything you approve.

A line with git in it asks if the repository names a program, whatever else is on it: when a line is not auto-allowed (a stage your own rule allowed, say echo, beside git status) it runs as typed, without the auto-allow environment, so the repository's core.fsmonitor, hooksPath, sshCommand, pager and gpg program count as well, and the question names the key.

A command outside the subset is still cut into its parts, so a deny or ask rule for rm reaches git status; rm x, a redirect to a file is shown as a write to its target, and the question names the part it is asking about. The cut is for the question and for deny and ask rules. Nothing is allowed because of it: a command outside the subset is allowed only by a bare bash allow rule or "fallback": "allow". Inside the subset a command the list above does not govern (make, rm) is for your rules, one stage at a time.

Trusting go test for your own repositories. Put the rule in your user config, ~/.config/dax/config.json, never a repository's (a project file's allow rules are ignored):

{ "policy": { "allow": ["bash(go test:*)", "bash(go vet:*)", "bash(go build:*)"] } }

It applies to every repository you open, so only add what you would run by hand in any of them. Each call is still one simple command: go test ./... && rm -rf x asks, as does go test ./... > out, and an arbitrary -exec, -toolexec or -vettool is yours to decide: tighten with "ask": ["bash(go test -exec:*)", "bash(go build -toolexec:*)", "bash(go vet -vettool:*)"] (the -o and -coverprofile flags write files and are worth asking about too).

Your rules in policy add to the default. Because deny beats ask beats allow, allow what the default asks about ("allow": ["write(docs/**)"]) and ask or deny what it allows. An ask the extensions ship, such as the one for a secret-looking path, holds against a plain allow of yours: lift it with a carve-out ("ask": ["deploy(!staging)"]) or, for the file tools, an allow that names the path ("allow": ["read(.env)"]). "builtin": false drops the shipped allow rules, so only your rules allow anything; the shipped asks and denies stay. "fallback": "allow" or "deny" changes what a call no rule names does.

Path rules (read(.env), write(docs/**), Read(secrets/**)) are written relative to the working directory and match the path after cleaning, so ./.env, a/../.env and the absolute path all meet the rule for .env, and docs/../.git/x is .git/x. Links are not followed for matching; the tools' own confinement still refuses one that leaves. glob, grep and ls are matched on the directory they search (default .), not on their pattern, and a rule for a directory should name dir and dir/**. A path rule is not a read ACL for a search that includes the directory from above.

An asked call prints ? allow bash {"command":"..."} (reason) [y/N]. Your answer is recorded in the session as a person's.

Security model

dax gives a model the ability to read your files, change them and run commands, so what it does and does not stand between the model and your machine matters.

What dax does

• Asks by default. Every call that is not on a short allow list asks you first: writes, edits, every command that is not one simple read-only command, memory writes and every MCP tool. Your answer is recorded.
• Auto-allows a small, checked set. The read-only tools (read, glob, grep, ls, confined to the working directory) and bash lines that parse in a safe subset of the syntax (simple commands joined by && and |, with a few trailing redirects) and whose every part is a read-only git, ls, cat/head/tail/wc/grep on named files, pwd, go version/go env NAME or cd into the workspace, with arguments that pass a per-command check and paths that stay inside the working directory (no .., links resolved). It is an allow-list of what is safe, not a list of what is dangerous: anything dax does not recognise asks. git is asked what its config would run first. go test, go build and go vet run the repository's code and are not on it.
• Confines the file tools. Paths outside the working directory, .., and symbolic links that lead out are refused, by the operating system's rooted open rather than by string checks.
• Treats the repository as untrusted. Its .dax/config.json can only tighten the policy; it cannot choose the provider, model or endpoint, add instructions, skills, memory or MCP servers, or allow anything. Its "builtin": false drops only allow rules, never an ask or a deny. Its AGENTS.md, .dax/skills and .dax/config.json are read through the workspace, so one that is a symbolic link out of it is refused as a tool's read would be: an AGENTS.md or .dax/skills is left out and reported, and a .dax/config.json is an error. A .dax/skills holding a link that leads outside it, even into the workspace, is left out too, since the skill tool reads without asking. Path rules match the path after normalisation, so docs/../.git/x is not under docs/**.
• Keeps credentials away from what it starts. Bash commands and MCP servers get your environment without *_API_KEY, *_TOKEN, *_SECRET and the like unless your config names a variable. Keys are read from the environment or from your own api_key_command, and never printed.
• Bounds resource use. read scans at most 2 MiB a call, grep skips big and non-regular files, search patterns cannot run away, and git runs with the repository's fsmonitor, pager and diff programs switched off.
• Keeps its records private. The session store and memory are created 0700; terminal control sequences in tool output and model text are stripped.

What dax does not do

• There is no sandbox. A command you approve runs with all your privileges, and a command that is auto-allowed is only as safe as the checks above. bash reaches files outside the working directory (approved commands are not confined, and the read-only allow list checks paths but cannot see what a program does with them). If you need a boundary, run dax in a container or VM.
• A prompt injection can still ask. A file, a web page, an MCP tool's output or an AGENTS.md can tell the model what to do. dax makes the dangerous steps ask; it cannot make you read the question. Read what you approve, especially a compound command, a write to .git/hooks, .dax/, AGENTS.md or .github/, and anything that sends data out.
• Your own allow rules are yours. "allow": ["write"] lets a model write .git/hooks/pre-commit; bash(go test:*) runs a hostile repository's tests with your privileges. Prefer narrow rules, and put deny or ask rules for the sensitive paths beside them.
• Committed secrets are not caught. The secret-path asks fire when a file is named. git log -p, git show and git diff without a path, grep over the tree and git show HEAD~5 print what a repository holds, and a .env or key that was ever committed is in it. Keep secrets out of history (and rotate one that got in); dax cannot tell which lines of a diff are keys.
• -trust-skills trusts the skills in directories you named (~/.dax/skills and your config's skills_dirs) and never the repository's: a skill in .dax/skills is text from the repository, and its allowed-tools are withheld, so it cannot run anything unasked.
• The provider sees what the model reads. Files and command output go to the model's provider (OpenAI, OpenRouter and whichever upstream it routes to, Anthropic, Google, your Ollama host, or the openresponses server you named). Choose the provider with that in mind; a path rule is not a read ACL for a search that includes the directory from above.
• The sub-agents are governed like the parent: every call explore or task makes is decided by the same rules (your denies, the secret-path asks, path rules, the auto-allow list), and one that asks is put to you, from inside the sub-agent's run. A task can write, so with "fallback": "allow" it writes unasked, as the main agent would.
• Not covered: programs the user's own git config names (it is trusted), a race between dax checking a path and the command using it, credential-file path variables such as KUBECONFIG (they pass through to commands), and denial of service by a model that loops (use Ctrl-C).

The terminal client

On a terminal, dax opens the terminal client (agentconsole) over the same session: the conversation rendered from the session's record, with the run's live deltas on top. When it exits it leaves one line on the terminal, the command that resumes the session:

To resume this session: dax -resume 01a112e5-b4df-7081-bd11-baac746b29cc

Any warning, such as a store whose permissions were fixed, is printed before the client takes the screen, and dax waits for Enter, since the client uses the alternate screen. With -v dax also prints its start lines (provider and model, the session, the policy in force, the tools, any omitted: line) and waits for Enter on an omission too, and when the client exits it prints what it noted during the run (skill grants, compactions, denied calls) above the resume command.

Each of dax's tool calls is drawn as what it does, on one line, with a short body under it:

tool	the line	collapsed	Ctrl-O
edit	the path, the lines it adds and removes (+3 −1)	the change as a diff, up to 8 lines	the whole diff
write	the path and the content's lines		the content
bash	$ and the command's first line, its timeout if set	how many lines it printed; a failure's exit status and last 3 lines	the whole output and the exit status
read	the path and the lines asked for (a.go:120-199)	how many lines it read	the output
grep, glob, ls	the pattern or directory, where and how	the matches and files, the files or the entries counted	the output
explore, task	the brief's first line, fork and main model when set	the report's first 2 lines	the report

A call that failed shows its error in place of the body. A call waiting on a permission shows what it would do: an edit its diff, a command of several lines the rest of it. Any other tool (skill, memory, an MCP server's) is drawn as the client draws any call, its name and arguments.

key
Enter	send a prompt, or steer the run in flight
Ctrl-C	copy the selection when one is drawn; otherwise abort the run, quit when idle (a second one quits at once)
y / n	approve / refuse the permission asked; n then takes an optional reason and Enter
PgUp, PgDn, Ctrl-Up/Down, Ctrl-Home/End, mouse wheel	scroll the conversation
Ctrl-R / Ctrl-O	show reasoning / tool arguments and output in full: the selected row's, or with no row selected every row's
Ctrl-T	the tree of the session's branches; Enter views one, c continues from it
Ctrl-P, Ctrl-N, Ctrl-B, Tab	move over the rows, continue from a row, open its detail (policy decision, who decided, verification)
click	select the row under the pointer; a click on the selected row expands or collapses it
drag	select text anywhere on the screen; it stays drawn until the next key or press
Ctrl-/ or F1	list every key; Esc, q or Ctrl-/ goes back

A call the policy asks about is a permission: the panel shows the call and the reason, which is the policy's with what it is asking about added: the rule that fired, the secret-looking path, the git config key that names a program, the part of a command line. Everything that can ask is answered on screen; nothing reads standard input once the client has the terminal. Two things cannot be a permission, which a run that ended answers, since the call that needs the answer is still running; the client asks them as questions while the run goes, ahead of any permission:

• A call a sub-agent (explore or task) makes that the policy asks about: y allows it, n refuses it with an optional reason, which the sub-agent sees, and Esc goes back.
• A question a tool asks mid-call (MCP elicitation) that is yes or no. One that is a form or a page to visit has no screen in the client, and is cancelled.

A call stamped as auto-allowed that changed before it ran fails with "the command changed since it was allowed ... ask again", which shows in the transcript.

In the REPL

dax -front repl (the default when not on a terminal) is a line REPL with slash commands the terminal client does not have yet, since the client takes the input line itself. A line typed while a run is in flight steers it and lands before the next model call; /follow text queues a follow-up that runs once the model would have stopped; /abort aborts. Between runs: /model name, /think on|off, /tools, /session, /mcp add <name> <command> (tools are mcp__<name>__<tool>; a name has letters, digits, - and _, and no __), /mcp remove <label>, /quit. In the terminal client use the flags and the config for the model, thinking and MCP servers (-model, -think, mcp_servers), and Ctrl-C to quit; /compact does not exist in either (use -compact N).

Sessions

Every run is recorded in one content-addressed store, ~/.dax/sessions, for every project. Admin commands:

dax -list                          # sessions recorded for this directory
dax -resume <id>                   # continue one, answering any call an abort cut off
dax -verify <id>                   # rebuild every request and check its hash
dax -project <id> -out dir         # write one as a JSONL file (RFC 0001)
dax -import old.jsonl              # bring a JSONL session into the store
dax -gc pack                       # pack loose objects; -gc sweep also drops what no session needs
dax -repair <id>                   # rewrite a damaged log from what still reads

-list, -verify and -project open the store read-only, so they work beside a running dax. -sessions "" disables recording.

Other state lives in ~/.dax: AGENTS.md (read before the project's), skills/, memory/. A project's .dax/skills and its AGENTS.md files are read too, through the workspace. The AGENTS.md chain follows the convention (https://agents.md): from the repository's root, the nearest directory at or above the start that holds a .git, down to where dax starts, the nearest file last; a directory above the repository is not read. With no repository only the start directory's file is read. A file you want in every session whatever the repository, such as one in a directory above your checkouts, goes in your config's agents_md_global.

Building on dax

dax is a Go module as well as a command. cmd/dax is one line, dax.Main with no options; a program of your own is the same line with an extension:

func main() {
	os.Exit(dax.Main(context.Background(), os.Args[1:],
		dax.WithName("acme", ""),
		dax.WithExtension(extension.Extension{
			Name:         "acme-deploy",
			Tools:        deployTools, // func(extension.ToolEnv) []agenttool.Tool
			Matchers:     extension.FixedMatchers(map[string]agentpolicy.ToolMatcher{"deploy": {Match: agentpolicy.GlobMatcher("env")}}),
			Policy:       policy.Rules{Allow: []string{"deploy(staging)"}},
			Instructions: "Never deploy to prod unless the user asks.",
			Renderers:    func(string) toolview.Renderers { return toolview.Renderers{"deploy": deployRenderer{}} },
		})))
}

dax's own extensions are built the same way, from the same fields:

Extension	Package	What it offers
dax-coding	ext/coding	read, write, edit, glob, grep, ls and bash; the allow list, the secret-path asks, the matchers and the aliases (Bash, Read, Edit, Write) for them; the stamp that holds an auto-allowed bash call to the plan the policy approved
dax-agents	ext/agents	the explore and task sub-agents, given every extension's tools (the read-only ones for explore)
dax-skills	ext/skills	the skill tool and catalogue; grants under -trust-skills
dax-memory	ext/memory	the memory tools and block

dax-coding is always on; the others follow agents, skills and memory in the settings. WithoutExtension leaves one of the four out whatever the settings say (any other name is an error), and WithExtension adds the program's after them.

extension.Extension field	What it adds
Name	the extension's name; its rules are recorded as extension:<Name>
Tools	tools (agenttool.Tool), in order, for the main agent and for task
ReadOnly	the names of its tools that only look, which explore gets too; one annotated destructive is refused
Owns	names of tools it adds through Kit (skill, memory_save), so its rules may name them
Matchers	how a rule's specifier reads a call of one of its tools: deploy(staging) above. Built over the session's ToolEnv, so one that looks at files looks at the session's workspace; extension.FixedMatchers wraps ones that read only the arguments
Aliases	names a rule may use for several of its tools: Read for read, grep, glob and ls
Policy	the allow, ask and deny rules it ships for its tools
Lifts	its tools whose asks a user's allow rule with a specifier lifts: read(.env)
BeforeToolCall	a hook, built over the session's ToolEnv, that decides or rewrites a call, folded with the policy, for the main agent and the sub-agents; extension.FixedHook wraps one that needs nothing of the session
Instructions	text in the main agent's system prompt, in extension order, before your instructions_file
Kit	agentkit options for anything else (skills, memory, child agents, guards), given every extension's tools
Renderers	how the terminal client draws its tools' calls (toolview.Renderer)

A session builds its extensions in two phases: every extension's tools, then each one's Kit options, so a sub-agent built in the second phase gets a tool an extension listed after it adds. Names are checked across the whole session, without regard to case: two extensions may not offer a tool, an alias or an owned name that differ only in case, mcp__ belongs to MCP servers, and an alias may differ from its own tool only in case (Bash and bash). The session's own kit options (the name, the model, the instructions, the policy, the session) go after the extensions', so where an option replaces another the session's is the one in force; a Kit that sets a policy on a session with the policy off is an error. The session closes the tools it built. One tool value is shared by every agent that has it, the main agent, explore and task, and may be called by them at once, so a tool that keeps state guards it.

An extension's tools are under the same policy as dax's. A call no rule allows asks first. Each extension's rules are a source of their own, extension:<Name>, which every verdict one of them decides names in the session's record, and the start line counts them. A rule may name only the extension's own tools and aliases, not a pattern and not a carve-out, so one extension cannot loosen another's tools or cancel its rules. Deny beats ask beats allow whatever the source: an extension's ask or deny holds against a plain allow in your config, and a carve-out in your config ("ask": ["deploy(!staging)"]) or, for a tool in Lifts, an allow with a specifier lifts it. A project's config can add asks and denies, and "builtin": false drops the extensions' allow rules but keeps their asks and denies.

An extension's tools get an extension.ToolEnv: the session's Workspace, Files over it, and MaxReadBytes. A tool that takes a path from the model reads and writes through Files (ReadFile, WriteFile, Update, Stat, ReadDir, Rel), which turns the path the model wrote, absolute in the workspace's root or relative to it, into the workspace's name and refuses one that leaves, as dax's file tools do. Update is a read-modify-write under the lock dax's write and edit take, so a tool's change and an edit beside it cannot lose each other's; a FIFO or a device is refused rather than waited on. A tool that runs a process uses Workspace.Exec, whose processes start with Workspace.Env(), the environment with credentials removed (a copy each call). Written that way, a tool runs unchanged wherever the session's workspace is. A tool is built over ToolEnv rather than agentkit's kit because what it needs is dax's, the workspace and its confinement, which the kit does not hold.

A renderer reads only the record. It declines (returns ok false) a call whose arguments or schema it was not written for, such as one an older version recorded, and the client draws that call as raw text. Two extensions that draw one tool are an error.

WithName puts the program's name in the banner, the resume command, the client header, the session header and the prompt; the config files and the store are still dax's (~/.config/dax, .dax, ~/.dax). The program keeps the flags, the config files, the providers, the session store, MCP, and the three fronts.

For a front of your own, build the session with agent.New and Options.Extensions, and drive its Turn (Session.Turn): prompt, steer, answer the permissions a run stops on (Turn.Permissions, Turn.Answer), reply to the questions asked while a call runs (Turn.Questions, Turn.Reply), and subscribe to the agent's events. Session is also agent.Controls (model, reasoning, MCP servers, Info) and Session.Record is the store to follow for a view. agent.Drive is a controller that answers by rule, for running without a person. The terminal client is agentconsole; hand it extension.Renderers(dir, exts). The public packages are dax, agent, extension, policy, tool, toolrender, workspace and the four under ext/; the configuration, providers, model metadata, prompt frame and renderer for the REPL stay internal. Everything is pre-1.0: the exported API may change in a minor version, and the changelog says when.

Where the tools act

A session acts in one workspace.Workspace: a root, a file system that never blocks on a FIFO or a device, writes and removes, an environment, one-shot execution with its output streamed as it arrives, and a descriptor the session records. workspace.Local is this machine's directory, opened as an os.Root so no name leaves it, with each command in its own process group. A container or a remote runtime is another implementation of the same interface, and the rest of dax cannot tell them apart:

• dax-coding's tools act through it: read, write, edit, glob, grep and ls through tool.Files, and bash through Workspace.Exec, with the same output wherever it runs.
• The policy's checks of a call read the same workspace: a path rule follows links through it, and the bash analyzer stats files, follows links, finds .git and reads git's configuration there (git is run through Exec, with paths passed as arguments). A workspace whose file system cannot tell a link from its target leaves those checks unable to follow links, and they ask.
• The project's instructions are read through it: the AGENTS.md chain (agentsmd's Options.FS), .dax/skills (an agentskill.Source over its file system) and .dax/config.json, each screened there, so a container gives its own and nothing comes from the directory dax started in. Only when the workspace is that directory on this machine and dax started below its repository's root are the AGENTS.md files between the two read from this machine, screened the same way.
• The session records the workspace: the header's and the env entry's cwd are its root, and the env entry names its kind (local, container, remote) and ref. The model is told the root as the working directory.

agent.Options.Workspace takes one; without it the session opens a workspace.Local over Dir and closes it. agent.Options.Store takes the session store the same way: an agentsession.Store the caller opened and closes, or, without it, the content-addressed store at Root. A remote store client (agentsession's RFC 0003) is meant to fit there. The interface follows the agentworkspace study in this workspace (examples/openhands-workspace), so dax moves to that module by a rename once it exists.

What is not there yet:

• dax ships only workspace.Local; a container or remote workspace is a program's to provide until the agentworkspace module exists.
• MCP servers run on this machine, with its environment scrubbed.
• RFC 0003's client is not a drop-in Store: opening takes a lease, an append returns a different result, and a lost lease needs handling. Options.Store takes today's interface; the lease handling waits for agentsession's remote client. Session.Path is empty for a store that is not local.
• A front over a wire (the terminal client driving a session elsewhere, ACP) is not built. The contract it would carry is the session's agent.Turn, which dax proposes for agentturn.

Develop

make check    # gofmt, go mod tidy -diff, go vet, staticcheck, govulncheck, go test -race
make build    # ./dax

In this workspace the module is built outside the go.work: GOWORK=off GOFLAGS=-mod=readonly make check. See AGENTS.md for the layout and release process, and CHANGELOG.md.

License

MIT. See LICENSE.

Documentation

Overview

Package dax is a coding agent: a terminal client, a REPL, or one prompt with -p, over an Open Responses model, under a policy, recording every session. Command dax (./cmd/dax) is Main with no options.

dax is a minimal core meant to be extended. The session (the model, the store, the policy, the workspace, AGENTS.md, MCP servers, compaction) knows no tool. Everything the model can do comes from an extension.Extension, and dax's own are extensions like any other: dax-coding (package ext/coding) has the file tools and bash, dax-agents the explore and task sub-agents, dax-skills skills, dax-memory memory. A program built on dax adds its own with WithExtension, on the same terms, and may leave one of dax's out with WithoutExtension; it keeps the flags, the config files, the providers, the session store and the fronts.

func main() {
	os.Exit(dax.Main(context.Background(), os.Args[1:],
		dax.WithName("acme", ""),
		dax.WithExtension(extension.Extension{
			Name:      "acme-deploy",
			Tools:     func(e extension.ToolEnv) []agenttool.Tool { return []agenttool.Tool{deploy.New(e.Workspace)} },
			Matchers:  extension.FixedMatchers(map[string]agentpolicy.ToolMatcher{"deploy": {Match: agentpolicy.GlobMatcher("env")}}),
			Policy:    policy.Rules{Allow: []string{"deploy(staging)"}},
			Renderers: func(string) toolview.Renderers { return toolview.Renderers{"deploy": deploy.Renderer{}} },
		})))
}

A program that wants a front of its own builds the session with agent.New and the same extensions.

dax is pre-1.0: the exported API of this package, agent, extension, policy, tool, toolrender and the ext packages may change in a minor version, and the CHANGELOG says when.

Example

A program built on dax: dax's flags, config, providers, session store, fronts and extensions, with one extension of its own holding two tools. changelog only looks, so the explore sub-agent has it too, and it runs unasked; deploy asks before every call, and the user's config can say otherwise with a rule such as deploy(staging), which the env matcher reads.

package main

import (
	"context"
	"encoding/json"
	"fmt"
	"os"

	"github.com/ChristopherDavenport/agentconsole/toolview"
	"github.com/ChristopherDavenport/agentconsole/view"
	"github.com/ChristopherDavenport/agentpolicy"
	"github.com/ChristopherDavenport/agenttool"

	"github.com/ChristopherDavenport/dax"
	"github.com/ChristopherDavenport/dax/extension"
	"github.com/ChristopherDavenport/dax/policy"
	"github.com/ChristopherDavenport/dax/workspace"
)

// changelogTool reads the project's CHANGELOG.md through the session's
// workspace, so it reads the same file whether the workspace is this
// machine, a container or a remote runtime, and reaches nothing outside.
func changelogTool(env extension.ToolEnv) []agenttool.Tool {
	return []agenttool.Tool{agenttool.NewFunc("changelog", "Read the project's changelog.", nil,
		func(context.Context, agenttool.Call) (agenttool.Result, error) {
			data, err := env.Files.ReadFile("CHANGELOG.md", env.MaxReadBytes)
			if err != nil {
				return agenttool.Result{}, err
			}
			return agenttool.Text(string(data)), nil
		})}
}

// deployTool is a tool that changes something outside the project.
func deployTool(env extension.ToolEnv) []agenttool.Tool {
	return []agenttool.Tool{agenttool.NewFunc("deploy", "Deploy the project to an environment.",
		json.RawMessage(`{"type":"object","properties":{"env":{"type":"string"}},"required":["env"]}`),
		func(ctx context.Context, c agenttool.Call) (agenttool.Result, error) {
			// The deploy runs where the session's tools act.
			out, err := env.Workspace.Exec(ctx, workspace.Command{Args: []string{"make", "deploy"}})
			if err != nil {
				return agenttool.Result{}, err
			}
			return agenttool.Text(fmt.Sprintf("deployed (exit %d)", out.ExitCode)), nil
		})}
}

// deployRenderer draws a deploy call in the terminal client.
type deployRenderer struct{}

func (deployRenderer) Head(c view.Call) (toolview.Line, bool) {
	var a struct{ Env string }
	if json.Unmarshal([]byte(c.Args), &a) != nil || a.Env == "" {
		return nil, false
	}
	return toolview.Line{toolview.S(toolview.Dim, "to "+a.Env)}, true
}

func (deployRenderer) Body(view.Call, bool) ([]toolview.Line, bool) { return nil, false }

// A program built on dax: dax's flags, config, providers, session store,
// fronts and extensions, with one extension of its own holding two tools. changelog only looks, so the
// explore sub-agent has it too, and it runs unasked; deploy asks before
// every call, and the user's config can say otherwise with a rule such
// as deploy(staging), which the env matcher reads.
func main() {
	os.Exit(dax.Main(context.Background(), os.Args[1:],
		dax.WithName("acme", ""),
		dax.WithExtension(extension.Extension{
			Name: "acme-release",
			Tools: func(e extension.ToolEnv) []agenttool.Tool {
				return append(changelogTool(e), deployTool(e)...)
			},
			ReadOnly:     []string{"changelog"},
			Matchers:     extension.FixedMatchers(map[string]agentpolicy.ToolMatcher{"deploy": {Match: agentpolicy.GlobMatcher("env")}}),
			Instructions: "Read the changelog before you describe a release. Never deploy to prod unless the user asks.",
			Policy:       policy.Rules{Allow: []string{"changelog"}},
			Renderers:    func(string) toolview.Renderers { return toolview.Renderers{"deploy": deployRenderer{}} },
		})))
}

Index

• func Main(ctx context.Context, args []string, opts ...Option) int
• type Option
  • func WithExtension(e extension.Extension) Option
  • func WithName(name, version string) Option
  • func WithoutExtension(name string) Option

Examples

• Package

Functions

func Main

func Main(ctx context.Context, args []string, opts ...Option) int

Main runs dax with args, the command line after the program's name, and returns the process's exit status: 0, or 1 after printing the error.

Types

type Option

type Option func(*program)

Option configures Main.

func WithExtension

func WithExtension(e extension.Extension) Option

WithExtension adds e after dax's extensions, in the order given. Two extensions that claim one name, a tool's, an alias or their own, are an error at start.

func WithName

func WithName(name, version string) Option

WithName names the program, for the usage line, the banner, the resume command, the client header and the session header, and for the agent in the system prompt. An empty version is the main module's, as `go install` stamps it, or devel for a build that has none; never dax's, which would put dax's version under another name. The config files and the session store are still dax's: ~/.config/dax, .dax and ~/.dax.

func WithoutExtension

func WithoutExtension(name string) Option

WithoutExtension leaves out the dax extension called name (dax-coding, dax-agents, dax-skills or dax-memory), whatever the settings say. A name that is none of them is an error at start.