agentainer 0.1.0 → 0.1.2

package/README.md

@@ -4,8 +4,8 @@ Run a team of coding agents — **Claude Code, Codex, Gemini CLI, Hermes** — s
 side in tmux, each in its own directory, each able to message the others only if
 your YAML file says it may.
 
-> Formerly **AgentSwarm**. Installed globally, the command is `agentainer`; the
-> repo's `./swarm.sh` remains a drop-in equivalent for local development.
+> Formerly **AgentSwarm**. Installed globally the command is `agentainer`; from a
+> clone, the `./agentainer` script in the repo root is the same thing.
 
 ```
    agents.yaml                 tmux
@@ -31,7 +31,7 @@ the agent's CLI, and types each agent's first prompt into it.
 
 - `tmux` (3.0+)
 - `python3` — PyYAML is used if present, otherwise a bundled parser handles the config
-- `node` (16+) — only for the global `agentainer` command; not needed if you run `./swarm.sh` from a clone
+- `node` (16+) — only for the global `agentainer` command; not needed if you run `./agentainer` from a clone
 - whichever agent CLIs you reference: `claude`, `codex`, `gemini`, `hermes` — install only the one(s) you actually use
 
 ## Install
@@ -51,7 +51,7 @@ Or from a clone (no npm needed):
 
 ```bash
 git clone https://github.com/mehmetcanfarsak/AgentSwarm.git && cd AgentSwarm
-./swarm.sh --help        # same commands as `agentainer`, straight from the repo
+./agentainer --help      # same commands as the global `agentainer`, straight from the repo
 ```
 
 ## Quickstart
@@ -69,15 +69,15 @@ agentainer down          # stop everything
 Give the swarm its actual work:
 
 ```bash
-./swarm.sh send --to orchestrator "Build a CLI that converts CSV to Parquet."
+agentainer send --to orchestrator "Build a CLI that converts CSV to Parquet."
 ```
 
 Watch the traffic between agents:
 
 ```bash
-./swarm.sh logs -f              # whole swarm, live
-./swarm.sh logs reviewer -n 20  # one agent
-./swarm.sh inbox developer      # messages an agent received
+agentainer logs -f              # whole swarm, live
+agentainer logs reviewer -n 20  # one agent
+agentainer inbox developer      # messages an agent received
 ```
 
 ---
@@ -135,15 +135,15 @@ Getting that to work reliably against a live TUI took more than a sleep:
 - **Claude Code silently discards keystrokes for several seconds partway through
   startup.** Measured on v2.1.205: input at t=2s landed, t=6s and t=12s vanished,
   t=20s landed. A fixed `boot_delay_ms` is therefore a coin flip. Before typing,
-  AgentSwarm types a throwaway token and waits for the input box to echo it back,
+  Agentainer types a throwaway token and waits for the input box to echo it back,
   then erases it (`ready_probe`). Enter is never sent, so nothing is submitted.
-- **Readiness is not monotonic**, so after pasting, AgentSwarm checks that the
+- **Readiness is not monotonic**, so after pasting, Agentainer checks that the
   text actually appeared on screen before pressing Enter, and retries if it did
   not. If delivery cannot be confirmed it refuses to press Enter, rather than
   submitting a half-delivered prompt.
 - **Both CLIs open a "do you trust this folder?" modal** on first run in a new
   directory, which would eat the first prompt (Enter answers the dialog). Claude
-  does this even under `--dangerously-skip-permissions`. AgentSwarm pre-trusts each
+  does this even under `--dangerously-skip-permissions`. Agentainer pre-trusts each
   agent's workdir: for codex in its generated `config.toml`, for claude by adding
   `hasTrustDialogAccepted` for that path in `~/.claude.json`.
 - **Both collapse a long paste into a chip** rather than showing the text —
@@ -199,7 +199,7 @@ tmux send-keys -t reviewer -l "your message" && tmux send-keys -t reviewer Enter
 **Agents get reminded when their answer goes nowhere.** A model that was asked a
 question will often just *write the answer as prose* and end its turn — and that
 prose reaches nobody, because only a `<swarm-send>` block is delivered. So when an
-agent owes a reply and finishes a turn without sending one, AgentSwarm messages it:
+agent owes a reply and finishes a turn without sending one, Agentainer messages it:
 
 ```
 Your last turn sent no message to anyone, and lead is waiting on your answer to
@@ -214,7 +214,7 @@ If instead the agent *tried* to send but the block was malformed, it gets the
 specific diagnosis — unclosed tag, missing `to`, unknown recipient, permission
 denied — so it can correct itself rather than lose the message silently.
 
-It is reminded at most `max_reply_reminders` times (default **1**), then AgentSwarm
+It is reminded at most `max_reply_reminders` times (default **1**), then Agentainer
 gives up and stops nagging. An agent that auto-forwards via `forward_responses_to`
 is never reminded, since its words did reach someone. Turn it off per agent with
 `reply_reminder: false`.
@@ -236,8 +236,8 @@ conversation instead of starting a fresh one — it does not re-send the first
 prompt, and it keeps any mail still queued for that agent:
 
 ```bash
-./swarm.sh sessions          # what is recorded, and the command that would resume it
-./swarm.sh up --resume       # reattach; agents without a recorded id start fresh
+agentainer sessions          # what is recorded, and the command that would resume it
+agentainer up --resume       # reattach; agents without a recorded id start fresh
 ```
 
 Claude is resumed with `--resume <id>`, codex with `resume <id>`. Set
@@ -262,7 +262,7 @@ agent sees on its own terminal. Use `can_talk_to: "*"` for "everyone else".
 
 ## Capturing what an agent says
 
-AgentSwarm needs to know when an agent finishes a turn — both to log it and to
+Agentainer needs to know when an agent finishes a turn — both to log it and to
 support auto-forwarding. How it finds out depends on the CLI, and the two
 mechanisms are **not** equally good:
 
@@ -276,7 +276,7 @@ mechanisms are **not** equally good:
   with **no `matcher` key** (`Stop` is not a tool event, and supplying one stops the
   interactive TUI from ever running the hook). It reads the session transcript.
   Claude fires the hook *before* flushing the assistant message to that transcript,
-  so AgentSwarm polls it briefly, and only reads text written after the last user
+  so Agentainer polls it briefly, and only reads text written after the last user
   message — otherwise a turn would silently capture nothing, or re-relay the
   previous turn's reply.
 - **codex** → the agent gets a private `CODEX_HOME` at `<agent-dir>/.codex/`
@@ -379,7 +379,7 @@ Some honest limits:
   busy forever. After `busy_timeout_ms` (default 15 minutes) it is treated as idle
   again, with a warning. `swarm idle <agent>` clears it immediately. Any mail queued
   for such an agent is not lost either: whenever some *other* agent finishes a turn,
-  AgentSwarm sweeps the now-idle agent's queue and delivers what was stranded — so
+  Agentainer sweeps the now-idle agent's queue and delivers what was stranded — so
   one missed turn-completion cannot wedge a queue permanently.
 - A capture only fires if the agent's `type` matches the CLI its `command` actually
   runs. If you point a `type: codex` agent at a `claude` command (e.g. through an
@@ -478,7 +478,7 @@ agent_types:
 
 `defaults:` supplies any agent key for agents that don't set it, including
 `workdir` — useful for putting a whole swarm in one repository. `templates:`
-overrides the text AgentSwarm generates — `comms` and `task_notice` (appended to
+overrides the text Agentainer generates — `comms` and `task_notice` (appended to
 first prompts), plus `reply_reminder` and `send_failed` (the nudges) —
 with `{agent} {swarm} {peers} {prefix} {inbox} {workdir}` available as
 placeholders.
@@ -497,9 +497,9 @@ Ready-to-run swarms in [`examples/`](examples/):
 | [`existing-repo.yaml`](examples/existing-repo.yaml) | Pairing | Two agents in one **existing** checkout, with `create_workdirs: false` |
 
 ```bash
-./swarm.sh validate -c examples/research-swarm.yaml   # look before you leap
-./swarm.sh up       -c examples/research-swarm.yaml
-./swarm.sh send --to lead "Research the state of WebGPU compute shaders."
+agentainer validate -c examples/research-swarm.yaml   # look before you leap
+agentainer up       -c examples/research-swarm.yaml
+agentainer send --to lead "Research the state of WebGPU compute shaders."
 ```
 
 `existing-repo.yaml` intentionally refuses to start until you point `workdir` at
@@ -509,28 +509,28 @@ a repository that exists.
 
 | Command | Purpose |
 |---|---|
-| `swarm.sh up` | Start the swarm. `--only a,b`, `--restart`, `--resume`, `--no-prompt`, `--attach` |
-| `swarm.sh down` | Kill sessions and watchers. `--only a,b` |
-| `swarm.sh restart` | `down` then `up` |
-| `swarm.sh status` | Table of agents, sessions, capture mode, permissions |
-| `swarm.sh attach <agent>` | Attach to an agent's tmux session |
-| `swarm.sh send --to <agent> "msg"` | Deliver a message (`--from`, `--file`, `--queue`, `--wait`, `--ignore-busy`, `--force`) |
-| `swarm.sh broadcast "msg"` | Message everyone the sender may talk to |
-| `swarm.sh sessions` | Show each agent's recorded conversation id (`--raw`) |
-| `swarm.sh queue <agent>` | Show what is waiting for a busy agent (`--clear`) |
-| `swarm.sh idle <agent>` | Force an agent back to idle, then drain its queue |
-| `swarm.sh inbox <agent>` | Print archived messages |
-| `swarm.sh logs [agent] [-f]` | Event log: prompts, responses, messages |
-| `swarm.sh validate` | Parse the config. `--show-prompts` renders final prompts |
-
-`./swarm.sh my-swarm.yaml` is shorthand for `./swarm.sh up -c my-swarm.yaml`.
+| `agentainer up` | Start the swarm. `--only a,b`, `--restart`, `--resume`, `--no-prompt`, `--attach` |
+| `agentainer down` | Kill sessions and watchers. `--only a,b` |
+| `agentainer restart` | `down` then `up` |
+| `agentainer status` | Table of agents, sessions, capture mode, permissions |
+| `agentainer attach <agent>` | Attach to an agent's tmux session |
+| `agentainer send --to <agent> "msg"` | Deliver a message (`--from`, `--file`, `--queue`, `--wait`, `--ignore-busy`, `--force`) |
+| `agentainer broadcast "msg"` | Message everyone the sender may talk to |
+| `agentainer sessions` | Show each agent's recorded conversation id (`--raw`) |
+| `agentainer queue <agent>` | Show what is waiting for a busy agent (`--clear`) |
+| `agentainer idle <agent>` | Force an agent back to idle, then drain its queue |
+| `agentainer inbox <agent>` | Print archived messages |
+| `agentainer logs [agent] [-f]` | Event log: prompts, responses, messages |
+| `agentainer validate` | Parse the config. `--show-prompts` renders final prompts |
+
+`agentainer my-swarm.yaml` is shorthand for `agentainer up -c my-swarm.yaml`.
 `-c` and `$SWARM_CONFIG` both select a config; `-c` wins.
 
 ## Layout
 
 ```
 AgentSwarm/
-├── swarm.sh                # entrypoint
+├── agentainer                # entrypoint
 ├── agents.example.yaml     # annotated config
 ├── llms.txt                # reference for agents configuring this tool
 ├── hooks/
@@ -568,7 +568,7 @@ being skipped, and a transcript read before Claude has flushed it.
 ## Troubleshooting
 
 **"could not confirm the text arrived; NOT pressing Enter".** The agent's input
-box never echoed the prompt, so AgentSwarm refused to submit it. Attach to the
+box never echoed the prompt, so Agentainer refused to submit it. Attach to the
 session to see what state the CLI is in -- usually a modal (login, trust,
 onboarding) is holding focus. Raise `ready_timeout_ms` if the CLI is merely slow.
 
@@ -587,7 +587,7 @@ exists and `.swarm/logs/hooks.log` for errors. If the agent also looks busy fore
 its `type` probably does not match the CLI its `command` runs (see the capture note
 under [Busy agents](#busy-agents-and-backpressure)).
 
-**Can't scroll up in an attached session.** AgentSwarm raises tmux's scrollback to
+**Can't scroll up in an attached session.** Agentainer raises tmux's scrollback to
 `tmux_history_limit` (50000 lines) and turns on `tmux_mouse`, so the wheel scrolls
 the backlog; press `q` to leave copy mode. If your terminal grabs the wheel itself,
 use `Ctrl-b [` then PageUp. Both options are set on the tmux server before sessions

package/{swarm.sh → agentainer}

@@ -1,14 +1,14 @@
 #!/usr/bin/env bash
 #
-# AgentSwarm -- launch a configurable swarm of coding agents in tmux.
+# Agentainer -- launch a configurable swarm of coding agents in tmux.
 #
-#   ./swarm.sh up                    start every agent in agents.yaml
-#   ./swarm.sh up -c my-swarm.yaml   ...from a different config
-#   ./swarm.sh agents.yaml           shorthand for `up -c agents.yaml`
-#   ./swarm.sh status                see who is running
-#   ./swarm.sh send --to dev "hi"    message an agent
-#   ./swarm.sh attach dev            jump into an agent's tmux session
-#   ./swarm.sh down                  stop everything
+#   ./agentainer up                    start every agent in agents.yaml
+#   ./agentainer up -c my-swarm.yaml   ...from a different config
+#   ./agentainer agents.yaml           shorthand for `up -c agents.yaml`
+#   ./agentainer status                see who is running
+#   ./agentainer send --to dev "hi"    message an agent
+#   ./agentainer attach dev            jump into an agent's tmux session
+#   ./agentainer down                  stop everything
 #
 # See README.md for the full reference, or llms.txt if you are an LLM.
 set -euo pipefail
@@ -27,7 +27,7 @@ if [[ -z "$PYTHON" ]]; then
 fi
 
 if [[ -z "$PYTHON" ]]; then
-  echo "xx AgentSwarm needs python3 on PATH (or set SWARM_PYTHON)" >&2
+  echo "xx Agentainer needs python3 on PATH (or set SWARM_PYTHON)" >&2
   exit 1
 fi
 

package/agents.example.yaml

@@ -1,9 +1,9 @@
 # =============================================================================
-# AgentSwarm example configuration
+# Agentainer example configuration
 #
 #   cp agents.example.yaml agents.yaml
-#   ./swarm.sh validate --show-prompts     # check it without launching anything
-#   ./swarm.sh up                          # launch the swarm
+#   agentainer validate --show-prompts     # check it without launching anything
+#   agentainer up                          # launch the swarm
 #
 # Every key is documented inline. Anything marked (optional) can be omitted.
 # =============================================================================
@@ -32,7 +32,7 @@ swarm:
   # (optional) How long to keep probing an agent's input box for a response
   # before giving up and typing the first prompt anyway. Some CLIs (Claude Code)
   # silently discard keystrokes for several seconds partway through startup, so
-  # AgentSwarm waits for the input box to echo a throwaway token before typing.
+  # Agentainer waits for the input box to echo a throwaway token before typing.
   ready_timeout_ms: 60000
 
   # (optional) An agent is "busy" from when a message is submitted to it until its
@@ -50,7 +50,7 @@ swarm:
 
   # (optional) An agent that answers a question as plain prose sends nothing: only a
   # <swarm-send> block is delivered. When it owes a reply and its turn ends without
-  # one, AgentSwarm messages it explaining how to send. This is how many times.
+  # one, Agentainer messages it explaining how to send. This is how many times.
   max_reply_reminders: 1
 
   # (optional) Reattach agents to their previous conversations on `up`, using the ids
@@ -63,9 +63,9 @@ swarm:
   pane_poll_ms: 700       # how often the watcher samples the pane
   pane_scrollback: 400    # lines of scrollback to diff
 
-  # (optional) tmux comfort settings for when you `swarm attach <agent>`.
+  # (optional) tmux comfort settings for when you `agentainer attach <agent>`.
   # The default tmux scrollback (2000 lines) is too small to hold a long
-  # multi-agent conversation, so AgentSwarm raises it before creating sessions.
+  # multi-agent conversation, so Agentainer raises it before creating sessions.
   # mouse mode lets you wheel-scroll the backlog; press q to leave copy mode.
   # Both are applied to the tmux server, so they also affect your other sessions.
   tmux_history_limit: 50000   # lines of scrollback per agent pane (0 = leave tmux's default)
@@ -85,13 +85,13 @@ defaults:
 # agent_types (optional) -- override the built-in launch commands, or define
 # entirely new agent types. Built-ins are: claude, codex, gemini, hermes.
 #
-#   capture: how AgentSwarm learns that an agent finished a turn.
+#   capture: how Agentainer learns that an agent finished a turn.
 #     hook -- the CLI can run an external program on turn completion.
 #             Supported for claude (Stop hook) and codex (notify program).
 #     pane -- no such facility: poll the tmux pane and diff it. Heuristic.
 #     none -- do not capture at all.
 #
-#   boot_delay_ms: a grace period before AgentSwarm starts probing the input
+#   boot_delay_ms: a grace period before Agentainer starts probing the input
 #     box. It does NOT need to cover the full startup time -- delivery of the
 #     first prompt is verified and retried, so a small value is fine.
 # -----------------------------------------------------------------------------
@@ -214,7 +214,7 @@ agents:
     # single `defaults.workdir` can serve every agent, e.g. "{root}/{name}-wt".
     #
     # Several agents may share one directory (they then see each other's edits;
-    # AgentSwarm warns you). Pair `create_workdir: false` with an existing
+    # Agentainer warns you). Pair `create_workdir: false` with an existing
     # project so a bad path is an error, not a new empty folder.
     #
     # workdir: ~/projects/acme-api
@@ -245,7 +245,7 @@ agents:
       You are the SCRIBE. Keep a running changelog of the project in NOTES.md.
 
 # -----------------------------------------------------------------------------
-# templates (optional) -- override the text AgentSwarm appends to first prompts.
+# templates (optional) -- override the text Agentainer appends to first prompts.
 # Available placeholders: {agent} {swarm} {peers} {prefix} {inbox} {workdir}
 # -----------------------------------------------------------------------------
 #

package/bin/agentainer.js

@@ -21,7 +21,7 @@ if (process.argv[2] === "doctor") {
   process.exit(r.status === null ? 1 : r.status);
 }
 
-// Locate a Python interpreter, mirroring swarm.sh (SWARM_PYTHON, then python3,
+// Locate a Python interpreter, mirroring ./agentainer (SWARM_PYTHON, then python3,
 // then python).
 function findPython() {
   const candidates = process.env.SWARM_PYTHON

package/examples/bug-hunt.yaml

@@ -1,8 +1,8 @@
 # =============================================================================
 # Bug hunt -- an automatic pipeline: reproduce -> diagnose -> fix -> verify.
 #
-#   ./swarm.sh up -c examples/bug-hunt.yaml
-#   ./swarm.sh send --to reproducer "Uploads over 2MB fail with a 500 in prod."
+#   agentainer up -c examples/bug-hunt.yaml
+#   agentainer send --to reproducer "Uploads over 2MB fail with a 500 in prod."
 #
 # This is the one shape where `forward_responses_to` earns its keep: each stage
 # hands its finished turn to the next stage automatically, with no human in the

package/examples/existing-repo.yaml

@@ -6,9 +6,9 @@
 # writes the code and a navigator who reads it as it lands.
 #
 #   1. Point `workdir` below at a real repository.
-#   2. ./swarm.sh validate -c examples/existing-repo.yaml
-#   3. ./swarm.sh up       -c examples/existing-repo.yaml
-#   4. ./swarm.sh send --to driver "Add retry-with-backoff to the HTTP client."
+#   2. agentainer validate -c examples/existing-repo.yaml
+#   3. agentainer up       -c examples/existing-repo.yaml
+#   4. agentainer send --to driver "Add retry-with-backoff to the HTTP client."
 #
 # Until you edit the paths, `validate` will refuse to run and tell you so --
 # that is the point of `create_workdir: false`.

package/examples/research-swarm.yaml

@@ -1,8 +1,8 @@
 # =============================================================================
 # Research swarm -- a small team that investigates a topic and writes it up.
 #
-#   ./swarm.sh up -c examples/research-swarm.yaml
-#   ./swarm.sh send --to lead "Research the state of WebGPU compute shaders."
+#   agentainer up -c examples/research-swarm.yaml
+#   agentainer send --to lead "Research the state of WebGPU compute shaders."
 #
 # Shape: hub and spoke. The lead hands out work and assembles the result; the
 # specialists only ever report back to the lead (and to each other where it
@@ -21,7 +21,7 @@ agents:
 
   - name: lead
     type: claude
-    command: "bash -ic 'chy3'"
+    command: "claude --dangerously-skip-permissions"
     can_talk_to: ["scout", "analyst", "writer"]
 
     # The lead waits for the human to supply the research question.
@@ -51,8 +51,8 @@ agents:
       Never accept "it depends" as a final answer. Force a recommendation.
 
   - name: scout
-    type: claude
-    command: "bash -ic 'chy3'"
+    type: gemini
+    command: "gemini --yolo"
     can_talk_to: ["lead", "analyst"]
 
     # Gemini has no turn-completion hook, so its output would have to be scraped
@@ -79,8 +79,8 @@ agents:
       of guessing. Say plainly when you could not find something.
 
   - name: analyst
-    type: claude
-    command: "bash -ic 'chy3'"
+    type: codex
+    command: "codex --yolo"
     can_talk_to: ["lead", "scout"]
 
     first_prompt: |
@@ -101,7 +101,7 @@ agents:
 
   - name: writer
     type: claude
-    command: "bash -ic 'chy3'"
+    command: "claude --dangerously-skip-permissions"
     can_talk_to: ["lead"]
 
     # A custom folder: the report lands somewhere predictable, outside the

package/examples/software-company.yaml

@@ -2,8 +2,8 @@
 # Software company -- a product team with a CTO, an architect, two developers,
 # a QA reviewer and a technical writer.
 #
-#   ./swarm.sh up -c examples/software-company.yaml
-#   ./swarm.sh send --to cto "Build a URL shortener with an API and a web UI."
+#   agentainer up -c examples/software-company.yaml
+#   agentainer send --to cto "Build a URL shortener with an API and a web UI."
 #
 # The communication graph is deliberately not a free-for-all: developers talk to
 # their architect and to QA, but not to each other, so design decisions go

package/hooks/claude_stop.sh

@@ -1,7 +1,7 @@
 #!/usr/bin/env bash
 # Claude Code `Stop` hook: fires when Claude finishes responding.
 # Claude passes a JSON payload on stdin containing `transcript_path`.
-# Installed automatically into <agent-workdir>/.claude/settings.json by `swarm up`.
+# Installed automatically into <agent-workdir>/.claude/settings.json by `agentainer up`.
 #
 # A hook must never break the agent it is attached to, so every failure here is
 # swallowed and the script always exits 0.
@@ -13,5 +13,5 @@ if [[ -n "${SWARM_ROOT:-}" ]] && mkdir -p "$SWARM_ROOT/.swarm/logs" 2>/dev/null;
   log="$SWARM_ROOT/.swarm/logs/hooks.log"
 fi
 
-"$HERE/swarm.sh" hook claude >>"$log" 2>&1
+"$HERE/agentainer" hook claude >>"$log" 2>&1
 exit 0

package/hooks/codex_notify.sh

@@ -12,5 +12,5 @@ if [[ -n "${SWARM_ROOT:-}" ]] && mkdir -p "$SWARM_ROOT/.swarm/logs" 2>/dev/null;
   log="$SWARM_ROOT/.swarm/logs/hooks.log"
 fi
 
-"$HERE/swarm.sh" hook codex "${1:-}" >>"$log" 2>&1
+"$HERE/agentainer" hook codex "${1:-}" >>"$log" 2>&1
 exit 0

package/lib/config.py

@@ -1,4 +1,4 @@
-"""Load, normalise and validate an AgentSwarm YAML config."""
+"""Load, normalise and validate an Agentainer YAML config."""
 
 from __future__ import annotations
 
@@ -70,9 +70,9 @@ NAME_RE = re.compile(r"^[A-Za-z0-9_][A-Za-z0-9_-]*$")
 
 DEFAULT_COMMS_TEMPLATE = """\
 ---
-## Swarm communication protocol
+## Inter-agent communication protocol
 
-You are the agent **{agent}** in the "{swarm}" swarm.
+You are the agent **{agent}** on the **{swarm}** team.
 Agents you are allowed to message: {peers}
 
 ### Receiving
@@ -108,7 +108,7 @@ literally, across as many lines as you need:
 To reach everyone you are allowed to talk to, use `<swarm-broadcast>` with no `to`.
 
 The `reply-to` attribute is optional; drop it when you are starting a new thread.
-Quoting it also tells the swarm your message is an answer, so the other agent is not
+Quoting it also marks your message as an answer, so the other agent is not
 chased for a reply to it. For an announcement that needs no answer, either use
 `<swarm-broadcast>` or add `expects-reply="false"`.
 
@@ -312,7 +312,7 @@ def load(path: str | os.PathLike) -> SwarmConfig:
         raise ConfigError(
             f"config file not found: {cfg_path}\n"
             "   Create one with:  cp agents.example.yaml agents.yaml\n"
-            "   Or point at it:   swarm.sh -c /path/to/swarm.yaml up"
+            "   Or point at it:   agentainer -c /path/to/swarm.yaml up"
         )
 
     try:
@@ -577,7 +577,7 @@ def load(path: str | os.PathLike) -> SwarmConfig:
         if not agent.workdir.exists() and not agent.create_workdir:
             raise ConfigError(
                 f"agent {agent.name!r}: workdir does not exist: {agent.workdir}\n"
-                "   Create it yourself, or allow AgentSwarm to: create_workdir: true"
+                "   Create it yourself, or allow Agentainer to: create_workdir: true"
             )
 
     shared: dict[Path, list[str]] = {}

package/lib/minyaml.py

@@ -1,7 +1,7 @@
 """A small YAML-subset parser.
 
-Used only when PyYAML is not importable, so that AgentSwarm keeps working on a
-bare Python 3 install. Supports the subset AgentSwarm configs actually need:
+Used only when PyYAML is not importable, so that Agentainer keeps working on a
+bare Python 3 install. Supports the subset Agentainer configs actually need:
 
   * nested block mappings and block sequences
   * scalars: strings, ints, floats, booleans, null

package/lib/swarm.py

@@ -1,7 +1,8 @@
 #!/usr/bin/env python3
-"""AgentSwarm -- run a swarm of coding agents in tmux and let them talk.
+"""Agentainer -- run a swarm of coding agents in tmux and let them talk.
 
-Invoked through ``swarm.sh``; see ``swarm.sh --help`` and README.md.
+Invoked through ``agentainer`` (or ``./agentainer`` from a clone); see
+``agentainer --help`` and README.md.
 """
 
 from __future__ import annotations
@@ -550,8 +551,8 @@ def read_sessions(cfg: SwarmConfig) -> dict:
 def write_sessions(cfg: SwarmConfig, agents: dict) -> None:
     cfg.runtime.mkdir(parents=True, exist_ok=True)
     header = (
-        "# AgentSwarm session state -- written automatically as agents work.\n"
-        "# `swarm up --resume` reads this to reattach each agent to its own\n"
+        "# Agentainer session state -- written automatically as agents work.\n"
+        "# `agentainer up --resume` reads this to reattach each agent to its own\n"
         "# conversation after a restart. Safe to delete; you then start fresh.\n"
     )
     body = yaml_dump(
@@ -847,7 +848,7 @@ def deliver(
     if not session_exists(target.session):
         raise SwarmError(
             f"agent {recipient!r} is not running (tmux session {target.session!r} missing). "
-            "Start it with: swarm up"
+            "Start it with: agentainer up"
         )
 
     msg_id = new_message_id()
@@ -872,7 +873,7 @@ def deliver(
         if not _paste_locked(cfg, target.session, body, enter=True, needle=needle):
             raise SwarmError(
                 f"could not confirm the message reached {recipient!r}; "
-                f"inspect it with: swarm attach {recipient}"
+                f"inspect it with: agentainer attach {recipient}"
             )
 
         with file_lock(cfg, recipient, "turn.lock"):
@@ -1231,7 +1232,7 @@ def install_codex_hook(agent: Agent) -> Path:
     # to that table. `notify` must therefore come before anything else, or codex
     # reads it as projects.<dir>.notify and never calls it.
     chunks = [
-        "# installed by AgentSwarm -- fires when codex finishes a turn.",
+        "# installed by Agentainer -- fires when codex finishes a turn.",
         "# Keep `notify` above every [table] header: TOML is order-sensitive.",
         f"notify = [{notify}]",
         "",
@@ -1609,8 +1610,8 @@ def write_shim(cfg: SwarmConfig) -> None:
     shim = cfg.bin_dir / "swarm"
     shim.write_text(
         "#!/usr/bin/env bash\n"
-        "# Generated by AgentSwarm. Lets an agent run `swarm send ...` from its shell.\n"
-        f'exec {shlex.quote(str(SWARM_HOME / "swarm.sh"))} "$@"\n'
+        "# Generated by Agentainer. Lets an agent run `swarm send ...` from its shell.\n"
+        f'exec {shlex.quote(str(SWARM_HOME / "agentainer"))} "$@"\n'
     )
     shim.chmod(0o755)
 
@@ -1803,7 +1804,7 @@ def cmd_up(args) -> int:
     print()
     info(f"swarm {cfg.name!r} is up with {len(started)} agent(s)")
     info(f"attach with:  tmux attach -t {started[0].session}")
-    info(f"or:           {SWARM_HOME / 'swarm.sh'} attach {started[0].name}")
+    info(f"or:           {SWARM_HOME / 'agentainer'} attach {started[0].name}")
 
     if args.attach:
         os.execvp("tmux", ["tmux", "attach", "-t", started[0].session])
@@ -2042,7 +2043,7 @@ def cmd_inbox(args) -> int:
     cfg = cfgmod.load(args.config)
     name = args.agent or os.environ.get("SWARM_AGENT")
     if not name:
-        die("specify an agent: swarm inbox <agent>")
+        die("specify an agent: agentainer inbox <agent>")
     cfg.get(name)
 
     box = cfg.inbox_dir / name
@@ -2129,7 +2130,7 @@ def select_agents(cfg: SwarmConfig, only: str | None) -> list[Agent]:
 
 
 def default_config() -> str:
-    """SWARM_CONFIG, else ./agents.yaml, else the agents.yaml beside swarm.sh."""
+    """SWARM_CONFIG, else ./agents.yaml, else the agents.yaml beside agentainer."""
     from_env = os.environ.get("SWARM_CONFIG")
     if from_env:
         return from_env

package/llms.txt

@@ -1,17 +1,17 @@
-# AgentSwarm
+# Agentainer
 
-> AgentSwarm runs several coding-agent CLIs (Claude Code, Codex, Gemini, Hermes) at once.
+> Agentainer runs several coding-agent CLIs (Claude Code, Codex, Gemini, Hermes) at once.
 > Each agent gets its own directory and its own tmux session. A single YAML file defines
 > the agents, the command that launches each one, its first prompt, and — as a strict
 > whitelist — which other agents it is allowed to message.
 >
-> This file is the reference for LLMs asked to write or repair an AgentSwarm config.
+> This file is the reference for LLMs asked to write or repair an Agentainer config.
 > If you are configuring a swarm, everything you need is below. Verify your work with
-> `./swarm.sh validate --show-prompts`, which parses the config and launches nothing.
+> `agentainer validate --show-prompts`, which parses the config and launches nothing.
 
 ## Mental model
 
-- One YAML file describes the whole swarm. `./swarm.sh up` reads it and, for each agent:
+- One YAML file describes the whole swarm. `agentainer up` reads it and, for each agent:
   creates `<root>/<name>/`, installs a turn-completion hook in that folder, opens a tmux
   session named `<session_prefix><name>`, runs the agent's `command` inside the folder,
   waits for its input box to become responsive, then types the assembled first prompt in.
@@ -82,7 +82,7 @@ agents:
     append_agents_that_you_can_talk_to_prompt: true
     in_first_prompt_append_your_task_will_be_sent_in_the_next_prompt: false
 
-templates:                 # override the text AgentSwarm generates
+templates:                 # override the text Agentainer generates
   comms: |                 # placeholders: {agent} {swarm} {peers} {prefix} {inbox} {workdir}
     You are {agent}. You may message: {peers}.
   task_notice: |
@@ -169,7 +169,7 @@ prose reaches nobody -- only a `<swarm-send>` block is delivered. So:
   to write.
 - If B *tried* to send but the block was malformed -- unclosed, missing `to`, unknown
   recipient, permission denied -- it gets the `send_failed` template naming each fault.
-- At most `max_reply_reminders` nudges (default 1), then AgentSwarm gives up silently.
+- At most `max_reply_reminders` nudges (default 1), then Agentainer gives up silently.
 - Anything B delivered counts, including an auto-forward via `forward_responses_to`.
   Reminders are sent from `swarm`, so they never create a reply obligation themselves.
 - `reply_reminder: false` disables it per agent; it is forced off when the agent's
@@ -223,23 +223,23 @@ Final prompt order: `first_prompt`, then the comms block, then the task notice.
 ## Commands
 
 ```
-./swarm.sh up [--only a,b] [--restart] [--no-prompt] [--attach]
-./swarm.sh down [--only a,b]
-./swarm.sh restart
-./swarm.sh status
-./swarm.sh attach <agent>
-./swarm.sh send --to <agent> [--from <name>] [--file F] [--queue] [--wait]
+agentainer up [--only a,b] [--restart] [--no-prompt] [--attach]
+agentainer down [--only a,b]
+agentainer restart
+agentainer status
+agentainer attach <agent>
+agentainer send --to <agent> [--from <name>] [--file F] [--queue] [--wait]
                              [--wait-timeout S] [--ignore-busy] [--force] "message"
-./swarm.sh broadcast "message"
-./swarm.sh queue <agent> [--clear]
-./swarm.sh idle <agent> [--no-drain]
-./swarm.sh inbox <agent> [-n N]
-./swarm.sh logs [agent] [-n N] [-f]
-./swarm.sh validate [--show-prompts]
+agentainer broadcast "message"
+agentainer queue <agent> [--clear]
+agentainer idle <agent> [--no-drain]
+agentainer inbox <agent> [-n N]
+agentainer logs [agent] [-n N] [-f]
+agentainer validate [--show-prompts]
 ```
 
 Config resolution: `-c PATH`, else `$SWARM_CONFIG`, else `./agents.yaml`, else the
-`agents.yaml` beside `swarm.sh`. `./swarm.sh x.yaml` is shorthand for `up -c x.yaml`.
+`agents.yaml` beside `agentainer`. `agentainer x.yaml` is shorthand for `up -c x.yaml`.
 
 ## What an agent sees at runtime
 
@@ -387,7 +387,7 @@ agents:
   limit stops them. Prefer one-directional forwarding, or none.
 - `--dangerously-skip-permissions` / `--yolo` let agents run tools unsupervised. Point
   `root` at a disposable directory.
-- The first prompt is typed into a live TUI. If AgentSwarm reports "could not confirm
+- The first prompt is typed into a live TUI. If Agentainer reports "could not confirm
   the text arrived", the CLI is usually stuck behind a modal (login, trust, onboarding);
   attach to the session and look. Raise `ready_timeout_ms` if it is merely slow.
 - PyYAML is used when installed; otherwise a bundled subset parser handles the config.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "agentainer",
-  "version": "0.1.0",
+  "version": "0.1.2",
   "description": "Run a configurable swarm of coding-agent CLIs (Claude Code, Codex, Gemini) in tmux and let them talk to each other.",
   "keywords": [
     "agents",
@@ -32,7 +32,7 @@
     "lib/*.py",
     "hooks/*.sh",
     "scripts/check-deps.js",
-    "swarm.sh",
+    "agentainer",
     "examples/*.yaml",
     "agents.example.yaml",
     "README.md",