init

A command that copies the complete, editable Cantos system into a chosen project directory and then starts its initial setup.

In plain words
What is it for?
It is for creating Cantos files such as assistants, workflows, references, logs, project context, and configuration in a safe destination.
Why use it?
It turns Cantos from a temporary plugin into a project-owned system that can be edited, tracked in Git, and developed over time.

Command

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/ne11nn/cantos-plugin/init
Clone the repo
git clone --depth 1 https://github.com/ne11nn/cantos-plugin
Per session 36 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 959 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00036 $0.00959
Opus 5 $0.00018 $0.00479
Sonnet 5 $0.00007 $0.00192
Haiku 4.5 $0.00004 $0.00096

Measured 2d ago against content hash 178719069c67, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

init scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

commands/init.md · 37 lines

What it actually says

Scaffold the complete Cantos system from the plugin into a directory the user chooses, so it becomes a real, editable, git-tracked system that evolves over time.

  • Source (bundled, read-only): ${CLAUDE_PLUGIN_ROOT}/system/
  • Default destination: the user's current working directory (cwd).

Cantos is a whole repo layout, not a single file. Scaffolding overlays many top-level entries (illustrative, not exhaustive — use the live ls -A in step 1 as the authoritative set): CLAUDE.md, .assistants/, .claude/, archives/, context/, decisions/, logs/, projects/, references/, registry/, tools/, workflows/, package.json, README.md, LICENSE, THIRD_PARTY_NOTICES.md, .gitignore. Treat this as potentially destructive and protect the user's existing files.

Steps:

  1. Pick a safe destination — never silently overwrite.

    • List the top-level entries the source would write: ls -A "${CLAUDE_PLUGIN_ROOT}/system".
    • Check the destination for collisions with that list (compare names, including the dotfiles .gitignore / .assistants / .claude). Then:
      • Destination is empty (or contains only files the source does not write): proceed to step 2.
      • Any collision exists: STOP. Show the user the exact colliding paths and offer three choices, recommending the first:
        1. Scaffold into a fresh subdirectory instead (e.g. ./cantos/), so nothing existing is touched. This is the safe default — recommend it whenever the cwd is an existing project.
        2. Overwrite only the listed colliding paths in place — only after the user explicitly confirms the full list. Call out .gitignore, README.md, LICENSE, and package.json specifically, since overwriting .gitignore can later cause secrets to be committed.
        3. Abort.
      • Never overwrite a CLAUDE.md, .gitignore, or any other existing file without the user's explicit, itemized confirmation.
  2. Copy the system into the chosen destination, preserving structure and symlinks. With DEST as the confirmed directory (cwd or the subdir), prefer:

    rsync -a --exclude='.git' "${CLAUDE_PLUGIN_ROOT}/system/" "$DEST"/
    

    If rsync is unavailable, use cp -a "${CLAUDE_PLUGIN_ROOT}/system/." "$DEST"/ (this preserves the symlinks under .claude/agents/). After copying, confirm a sample agent symlink resolves, e.g. ls -lL "$DEST"/.claude/agents/browser-agent.md.

  3. Confirm the scaffold. List what landed in DEST: CLAUDE.md, .assistants/ (cantos, folio, lyren, pylon), .claude/ (skills, agents, rules, templates), workflows/, references/, registry/, context/ (placeholders), decisions/, tools/.

  4. Run first-run setup now. The scaffolded context/me.md still contains the literal marker <!-- SETUP-NOT-DONE -->. Read "$DEST"/workflows/cantos/first_run_setup.md and run it start to finish in this session: interview the user, personalize context/me.md and context/work.md, set the communication style, tune the starter assistants (rename, remove, or add), log the decision, then remove the marker. This is the same gate the scaffolded CLAUDE.md enforces on the first message; running it now completes setup immediately. (If the user scaffolded into a subdirectory, note that Cantos will auto-load only when Claude Code is started from inside DEST.)

  5. Tell the user the system is live. From the next session started in DEST, CLAUDE.md loads automatically as project memory and Claude Code begins as Cantos. Recommend they keep that directory in a private repo, since it will hold their personal context; the bundled README.md and its Privacy notes explain why.

Then proceed.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 37 lines · 36 tokens per session scan A 178719069c67

Subscribe to this mod's changes

init is a command published in the GitHub repository ne11nn/cantos-plugin (1 stars, last pushed 2d ago), licensed MIT. It adds 36 tokens to every session and 959 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.